Posts Tagged ‘imaging biomarkers’

Imaging-Biomarkers; from discovery to validation

Author: Dror Nir, PhD.


Recent technology advances such as miniaturization and improvement in electronic-processing components is driving increased introduction of innovative medical-imaging devices into critical nodes of major-diseases’ management pathways. Similarly, medical imaging bears outstanding potential to improve the process of drugs development and regulation (e.g. companion diagnostics and imaging surrogate markers. In; The Role of Medical Imaging in Personalized Medicine I discussed in length the role medical imaging assumes in drugs development.  Integrating imaging into drug development processes, specifically at the early stages of drug discovery, as well as for monitoring drug delivery and the response of targeted processes to the therapy is a growing trend. A nice (and short) review highlighting the processes, opportunities, and challenges of medical imaging in new drug development is: Medical imaging in new drug clinical development. An important aspect of drug development that is largely discussed is facilitating testing of the new drug through clinical studies. A major hurdle in development of many anti-cancer drugs is the long time that is required to determine the efficacy of the new drug through measurement of clinically meaningful endpoints; e.g. overall survival. Imaging is offering the opportunity to determine surrogate markers of clinical outcome (as a substitute for a clinically meaningful endpoints). The need for surrogate outcome markers is especially great with newer agents that may act by tumour stabilization as opposed to shrinkage.

To comply with current trends; e.g. personalized medicine and evidence-based medicine, medical imaging must support quantification of meaningful pathological phenomena; e.g. morphological deformations, enhanced/reduced chemical reactions, presence/absence of biological substances etc….


Two examples: 

Molecular imaging (e.g. PET, MRS) allows the visual representation, characterization, and quantification of biological processes at the cellular and subcellular levels within intact living organisms. In oncology, it can be used to depict the abnormal molecules as well as the aberrant interactions of altered molecules on which cancers depend. An established biological process is neoplastic angiogenesis is associated with a number of detectable changes at molecular and microcirculatory levels. In Positron emission tomographic imaging of angiogenesis and vascular function the authors are offering that direct study of angiogenic molecular biology and tumour circulation before during and after treatment may offer useful surrogate markers for vascular-targeted therapies. The paper reviews two main areas: (a) the methodology behind PET imaging of tumour blood supply with 15O-oxygen labelled compounds; and (b) newer tracers in development as markers of angiogenetic biology.

A largely sought-for application for medical imaging is Monitoring quality of surgery: Cancer patients could benefit from a surgical procedure that helps the surgeon to determine adequate tumor resection margins. Variety of applications and work-flows; e.g. Systemic injection of tumor-specific fluorescence agents with subsequent intraoperative optical imaging to guide the surgeon in the process are offered. Recently, in order to overcome the problem of tumor heterogeneity it was proposed to shift the focus of tumor targeting towards the follicle-stimulating hormone receptor (FSHR).

Imaging bio-markers

Being able to discover and clinically validate fundamental finger-prints of cancer which can be detected and quantified through medical-imaging modalities is key to transforming the potential presented by medical imaging into clinical reality. Such specific finger-prints/characteristics are usually referred to as imaging bio-markers.

A critical step in the discovery and validation of imaging bio-markers is the matching of tissue location as depicted by imaging-products (most commonly images) to their histology, as underlined by a pathologist under the microscope.

Since histology requires extraction of organ tissue and some processing, it is impossible to achieve such matching in real time. Therefore, different techniques were developed to support the retrospective matching between histology and imaging. The most prevalent one rely on image registration: i.e. the products of medical imaging are registered to images of pathology slides. The main limitation of such methods has to do with:

  1. The fact that the two images poses largely different image resolution.
  2. The form-factor (shape and dimensions) of Histological tissue-slides are distorted in comparison to their in-vivo state.
  3. Histology-reading is subjective; i.e. the concordance between readings of different pathologist is far from being satisfactory. It gets worse when it comes to staging of the cancer.
  4. There is large variation in the quality of medical imaging products.

A Workflow to Improve the Alignment of Prostate Imaging with Whole-mount Histopathology presents a robust methodology validating imaging biomarkers in the case of prostate cancer. In this paper we describe a workflow for three-dimensional alignment of prostate imaging data against whole-mount prostatectomy reference specimens and assess its performance against a standard workflow. We hypothesized that integration of image registration principles into the histological workflow for radical prostatectomy specimens would increase the alignment accuracy. In this post I will include only few excerpts from this paper which I strongly recommend to read in full.

Materials and Methods

Ethical approval was granted. Patients underwent motorized transrectal ultrasound (Prostate Histoscanning) to generate a three-dimensional image of the prostate before radical prostatectomy. The test workflow incorporated steps for axial alignment between imaging and histology, size adjustments following formalin fixation, and use of custom-made parallel cutters and digital caliper instruments. The control workflow comprised freehand cutting and assumed homogeneous block thicknesses at the same relative angles between pathology and imaging sections. The basic requirements of image registration were incorporated within the pathological protocol.

We demonstrate that the use of a simple, custom-made tissue-planer to slice the formalin-fixed prostate results in more uniform and parallel tissue blocks than conventional freehand techniques, and increases the accuracy of image alignment.  We also show that accounting for dimensional change due to formalin fixation is essential during image alignment.

Figure 1: Suggested workflow for registration of scanned histopathological data with radiological imaging


 Figure 3

A sketch of the tissue cutting device is shown (A).  The formalin-fixed prostate was placed on the space marked “X” on the device with its flat posterior surface facing down.  With the probe in the urethra to align the AP axis with the device, the base of the gland was gently pressed onto “Y”.  The probe was then removed, and a mounted microtome blade was lowered along the 4mm raised edge of the device from top to bottom to cut away the block (B).  The sliced block was put aside with its apical face facing down, and the process was repeated by gently pressing the cut surface flush against the device before each cut (C).  The thickness of each block was measured in 5 locations marked (D).



Thirty radical prostatectomy specimens were histologically and radiologically processed, either by an alignment-optimized workflow (n = 20) or a control workflow (n = 10). The optimized workflow generated tissue blocks of heterogeneous thicknesses but with no significant drifting in the cutting plane. The control workflow resulted in significantly nonparallel blocks, accurately matching only one out of four histology blocks to their respective imaging data. The image-to-histology alignment accuracy was 20% greater in the optimized workflow (P < .0001), with higher sensitivity (85% vs. 69%) and specificity (94% vs. 73%) for margin prediction in a 5 × 5-mm grid analysis.

Figure 5. Assessment of alignment accuracy between radiological images and pathological sections

The method of assessing alignment accuracy between radiological images and pathological slides is shown using an example.  Each square within the grids overlaid onto histology and radiological images were scored either as a “1”, indicating the presence of a histological or radiological margin, respectively, or “0”.  Scored pathology grids were used as the reference, and scored radiology grids were used as the index.  Hence, we determined true positives i.e. grid points score “1” in both histology and radiology (yellow squares, n=25), false positives i.e. grid points on the radiology scores “1” but not on histology (green squares, n=4), false negatives i.e. grid points on the histology scores “0” but not on radiology (red squares, n=3), and true negatives (grey squares, n=38).



A significantly better alignment was observed in the optimized workflow. Evaluation of prostate imaging biomarkers using whole-mount histology references should include a test-to-reference spatial alignment workflow.


Read Full Post »

The importance of spatially-localized and quantified image interpretation in cancer management

Writer & reporter: Dror Nir, PhD

I became involved in the development of quantified imaging-based tissue characterization more than a decade ago. From the start, it was clear to me that what clinicians needs will not be answered by just identifying whether a certain organ harbors cancer. If imaging devices are to play a significant role in future medicine, as a complementary source of information to bio-markers and gene sequencing the minimum value expected of them is accurate directing of biopsy needles and treatment tools to the malignant locations in the organ.  Therefore, the design goal of the first Prostate-HistoScanning (“PHS”) version I went into the trouble of characterizing localized volume of tissue at the level of approximately 0.1cc (1x1x1 mm). Thanks to that, the imaging-interpretation overlay of PHS localizes the suspicious lesions with accuracy of 5mm within the prostate gland; Detection, localisation and characterisation of prostate cancer by prostate HistoScanning(™).

I then started a more ambitious research aiming to explore the feasibility of identifying sub-structures within the cancer lesion itself. The preliminary results of this exploration were so promising that it surprised not only the clinicians I was working with but also myself. It seems, that using quality ultrasound, one can find Imaging-Biomarkers that allows differentiation of inside structures of a cancerous lesions. Unfortunately, for everyone involved in this work, including me, this scientific effort was interrupted by financial constrains before reaching maturity.

My short introduction was made to explain why I find the publication below important enough to post and bring to your attention.

I hope for your agreement on the matter.

Quantitative Imaging in Cancer Evolution and Ecology

Robert A. Gatenby, MD, Olya Grove, PhD and Robert J. Gillies, PhD

From the Departments of Radiology and Cancer Imaging and Metabolism, Moffitt Cancer Center, 12902 Magnolia Dr, Tampa, FL 33612. Address correspondence to  R.A.G. (e-mail:


Cancer therapy, even when highly targeted, typically fails because of the remarkable capacity of malignant cells to evolve effective adaptations. These evolutionary dynamics are both a cause and a consequence of cancer system heterogeneity at many scales, ranging from genetic properties of individual cells to large-scale imaging features. Tumors of the same organ and cell type can have remarkably diverse appearances in different patients. Furthermore, even within a single tumor, marked variations in imaging features, such as necrosis or contrast enhancement, are common. Similar spatial variations recently have been reported in genetic profiles. Radiologic heterogeneity within tumors is usually governed by variations in blood flow, whereas genetic heterogeneity is typically ascribed to random mutations. However, evolution within tumors, as in all living systems, is subject to Darwinian principles; thus, it is governed by predictable and reproducible interactions between environmental selection forces and cell phenotype (not genotype). This link between regional variations in environmental properties and cellular adaptive strategies may permit clinical imaging to be used to assess and monitor intratumoral evolution in individual patients. This approach is enabled by new methods that extract, report, and analyze quantitative, reproducible, and mineable clinical imaging data. However, most current quantitative metrics lack spatialness, expressing quantitative radiologic features as a single value for a region of interest encompassing the whole tumor. In contrast, spatially explicit image analysis recognizes that tumors are heterogeneous but not well mixed and defines regionally distinct habitats, some of which appear to harbor tumor populations that are more aggressive and less treatable than others. By identifying regional variations in key environmental selection forces and evidence of cellular adaptation, clinical imaging can enable us to define intratumoral Darwinian dynamics before and during therapy. Advances in image analysis will place clinical imaging in an increasingly central role in the development of evolution-based patient-specific cancer therapy.

© RSNA, 2013



Cancers are heterogeneous across a wide range of temporal and spatial scales. Morphologic heterogeneity between and within cancers is readily apparent in clinical imaging, and subjective descriptors of these differences, such as necrotic, spiculated, and enhancing, are common in the radiology lexicon. In the past several years, radiology research has increasingly focused on quantifying these imaging variations in an effort to understand their clinical and biologic implications (1,2). In parallel, technical advances now permit extensive molecular characterization of tumor cells in individual patients. This has led to increasing emphasis on personalized cancer therapy, in which treatment is based on the presence of specific molecular targets (3). However, recent studies (4,5) have shown that multiple genetic subpopulations coexist within cancers, reflecting extensive intratumoral somatic evolution. This heterogeneity is a clear barrier to therapy based on molecular targets, since the identified targets do not always represent the entire population of tumor cells in a patient (6,7). It is ironic that cancer, a disease extensively and primarily analyzed genetically, is also the most genetically flexible of all diseases and, therefore, least amenable to such an approach.

Genetic variations in tumors are typically ascribed to a mutator phenotype that generates new clones, some of which expand into large populations (8). However, although identification of genotypes is of substantial interest, it is insufficient for complete characterization of tumor dynamics because evolution is governed by the interactions of environmental selection forces with the phenotypic, not genotypic, properties of populations as shown, for example, by evolutionary convergence to identical phenotypes among cave fish even when they are from different species (911). This connection between tissue selection forces and cellular properties has the potential to provide a strong bridge between medical imaging and the cellular and molecular properties of cancers.

We postulate that differences within tumors at different spatial scales (ie, at the radiologic, cellular, and molecular [genetic] levels) are related. Tumor characteristics observable at clinical imaging reflect molecular-, cellular-, and tissue-level dynamics; thus, they may be useful in understanding the underlying evolving biology in individual patients. A challenge is that such mapping across spatial and temporal scales requires not only objective reproducible metrics for imaging features but also a theoretical construct that bridges those scales (Fig 1).


Figure 1a: Computed tomographic (CT) scan of right upper lobe lung cancer in a 50-year-old woman.


Figure 1b: Isoattenuation map shows regional heterogeneity at the tissue scale (measured in centimeters).


Figure 1c & 1d: (c, d)Whole-slide digital images (original magnification, ×3) of a histologic slice of the same tumor at the mesoscopic scale (measured in millimeters) (c) coupled with a masked image of regional morphologic differences showing spatial heterogeneity (d). 


Figure 1e: Subsegment of the whole slide image shows the microscopic scale (measured in micrometers) (original magnification, ×50).


Figure 1f: Pattern recognition masked image shows regional heterogeneity. In a, the CT image of non–small cell lung cancer can be analyzed to display gradients of attenuation, which reveals heterogeneous and spatially distinct environments (b). Histologic images in the same patient (c, e) reveal heterogeneities in tissue structure and density on the same scale as seen in the CT images. These images can be analyzed at much higher definition to identify differences in morphologies of individual cells (3), and these analyses reveal clusters of cells with similar morphologic features (d, f). An important goal of radiomics is to bridge radiologic data with cellular and molecular characteristics observed microscopically.

To promote the development and implementation of quantitative imaging methods, protocols, and software tools, the National Cancer Institute has established the Quantitative Imaging Network. One goal of this program is to identify reproducible quantifiable imaging features of tumors that will permit data mining and explicit examination of links between the imaging findings and the underlying molecular and cellular characteristics of the tumors. In the quest for more personalized cancer treatments, these quantitative radiologic features potentially represent nondestructive temporally and spatially variable predictive and prognostic biomarkers that readily can be obtained in each patient before, during, and after therapy.

Quantitative imaging requires computational technologies that can be used to reliably extract mineable data from radiographic images. This feature information can then be correlated with molecular and cellular properties by using bioinformatics methods. Most existing methods are agnostic and focus on statistical descriptions of existing data, without presupposing the existence of specific relationships. Although this is a valid approach, a more profound understanding of quantitative imaging information may be obtained with a theoretical hypothesis-driven framework. Such models use links between observable tumor characteristics and microenvironmental selection factors to make testable predictions about emergent phenotypes. One such theoretical framework is the developing paradigm of cancer as an ecologic and evolutionary process.

For decades, landscape ecologists have studied the effects of heterogeneity in physical features on interactions between populations of organisms and their environments, often by using observation and quantification of images at various scales (1214). We propose that analytic models of this type can easily be applied to radiologic studies of cancer to uncover underlying molecular, cellular, and microenvironmental drivers of tumor behavior and specifically, tumor adaptations and responses to therapy (15).

In this article, we review recent developments in quantitative imaging metrics and discuss how they correlate with underlying genetic data and clinical outcomes. We then introduce the concept of using ecology and evolutionary models for spatially explicit image analysis as an exciting potential avenue of investigation.


Quantitative Imaging and Radiomics

In patients with cancer, quantitative measurements are commonly limited to measurement of tumor size with one-dimensional (Response Evaluation Criteria in Solid Tumors [or RECIST]) or two-dimensional (World Health Organization) long-axis measurements (16). These measures do not reflect the complexity of tumor morphology or behavior, and in many cases, changes in these measures are not predictive of therapeutic benefit (17). In contrast, radiomics (18) is a high-throughput process in which a large number of shape, edge, and texture imaging features are extracted, quantified, and stored in databases in an objective, reproducible, and mineable form (Figs 12). Once transformed into a quantitative form, radiologic tumor properties can be linked to underlying genetic alterations (the field is called radiogenomics) (1921) and to medical outcomes (2227). Researchers are currently working to develop both a standardized lexicon to describe tumor features (28,29) and a standard method to convert these descriptors into quantitative mineable data (30,31) (Fig 3).


Figure 2: Contrast-enhanced CT scans show non–small cell lung cancer (left) and corresponding cluster map (right). Subregions within the tumor are identified by clustering pixels based on the attenuation of pixels and their cumulative standard deviation across the region. While the entire region of interest of the tumor, lacking the spatial information, yields a weighted mean attenuation of 859.5 HU with a large and skewed standard deviation of 243.64 HU, the identified subregions have vastly different statistics. Mean attenuation was 438.9 HU ± 45 in the blue subregion, 210.91 HU ± 79 in the yellow subregion, and 1077.6 HU ± 18 in the red subregion.



Figure 3: Chart shows the five processes in radiomics.

Several recent articles underscore the potential power of feature analysis. After manually extracting more than 100 CT image features, Segal and colleagues found that a subset of 14 features predicted 80% of the gene expression pattern in patients with hepatocellular carcinoma (21). A similar extraction of features from contrast agent–enhanced magnetic resonance (MR) images of glioblastoma was used to predict immunohistochemically identified protein expression patterns (22). Other radiomic features, such as texture, can be used to predict response to therapy in patients with renal cancer (32) and prognosis in those with metastatic colon cancer (33).

These pioneering studies were relatively small because the image analysis was performed manually, and the studies were consequently underpowered. Thus, recent work in radiomics has focused on technical developments that permit automated extraction of image features with the potential for high throughput. Such methods, which rely heavily on novel machine learning algorithms, can more completely cover the range of quantitative features that can describe tumor heterogeneity, such as texture, shape, or margin gradients or, importantly, different environments, or niches, within the tumors.

Generally speaking, texture in a biomedical image is quantified by identifying repeating patterns. Texture analyses fall into two broad categories based on the concepts of first- and second-order spatial statistics. First-order statistics are computed by using individual pixel values, and no relationships between neighboring pixels are assumed or evaluated. Texture analysis methods based on first-order statistics usually involve calculating cumulative statistics of pixel values and their histograms across the region of interest. Second-order statistics, on the other hand, are used to evaluate the likelihood of observing spatially correlated pixels (34). Hence, second-order texture analyses focus on the detection and quantification of nonrandom distributions of pixels throughout the region of interest.

The technical developments that permit second-order texture analysis in tumors by using regional enhancement patterns on dynamic contrast-enhanced MR images were reviewed recently (35). One such technique that is used to measure heterogeneity of contrast enhancement uses the Factor Analysis of Medical Image Sequences (or FAMIS) algorithm, which divides tumors into regions based on their patterns of enhancement (36). Factor Analysis of Medical Image Sequences–based analyses yielded better prognostic information when compared with region of interest–based methods in numerous cancer types (1921,3739), and they were a precursor to the Food and Drug Administration–approved three-time-point method (40). A number of additional promising methods have been developed. Rose and colleagues showed that a structured fractal-based approach to texture analysis improved differentiation between low- and high-grade brain cancers by orders of magnitude (41). Ahmed and colleagues used gray level co-occurrence matrix analyses of dynamic contrast-enhanced images to distinguish benign from malignant breast masses with high diagnostic accuracy (area under the receiver operating characteristic curve, 0.92) (26). Others have shown that Minkowski functional structured methods that convolve images with differently kernelled masks can be used to distinguish subtle differences in contrast enhancement patterns and can enable significant differentiation between treatment groups (42).

It is not surprising that analyses of heterogeneity in enhancement patterns can improve diagnosis and prognosis, as this heterogeneity is fundamentally based on perfusion deficits, which generate significant microenvironmental selection pressures. However, texture analysis is not limited to enhancement patterns. For example, measures of heterogeneity in diffusion-weighted MR images can reveal differences in cellular density in tumors, which can be matched to histologic findings (43). Measures of heterogeneity in T1- and T2-weighted images can be used to distinguish benign from malignant soft-tissue masses (23). CT-based texture features have been shown to be highly significant independent predictors of survival in patients with non–small cell lung cancer (24).

Texture analyses can also be applied to positron emission tomographic (PET) data, where they can provide information about metabolic heterogeneity (25,26). In a recent study, Nair and colleagues identified 14 quantitative PET imaging features that correlated with gene expression (19). This led to an association of metagene clusters to imaging features and yielded prognostic models with hazard ratios near 6. In a study of esophageal cancer, in which 38 quantitative features describing fluorodeoxyglucose uptake were extracted, measures of metabolic heterogeneity at baseline enabled prediction of response with significantly higher sensitivity than any whole region of interest standardized uptake value measurement (22). It is also notable that these extensive texture-based features are generally more reproducible than simple measures of the standardized uptake value (27), which can be highly variable in a clinical setting (44).


Spatially Explicit Analysis of Tumor Heterogeneity

Although radiomic analyses have shown high prognostic power, they are not inherently spatially explicit. Quantitative border, shape, and texture features are typically generated over a region of interest that comprises the entire tumor (45). This approach implicitly assumes that tumors are heterogeneous but well mixed. However, spatially explicit subregions of cancers are readily apparent on contrast-enhanced MR or CT images, as perfusion can vary markedly within the tumor, even over short distances, with changes in tumor cell density and necrosis.

An example is shown in Figure 2, which shows a contrast-enhanced CT scan of non–small cell lung cancer. Note that there are many subregions within this tumor that can be identified with attenuation gradient (attenuation per centimeter) edge detection algorithms. Each subregion has a characteristic quantitative attenuation, with a narrow standard deviation, whereas the mean attenuation over the entire region of interest is a weighted average of the values across all subregions, with a correspondingly large and skewed distribution. We contend that these subregions represent distinct habitats within the tumor, each with a distinct set of environmental selection forces.

These observations, along with the recent identification of regional variations in the genetic properties of tumor cells, indicate the need to abandon the conceptual model of cancers as bounded organlike structures. Rather than a single self-organized system, cancers represent a patchwork of habitats, each with a unique set of environmental selection forces and cellular evolution strategies. For example, regions of the tumor that are poorly perfused can be populated by only those cells that are well adapted to low-oxygen, low-glucose, and high-acid environmental conditions. Such adaptive responses to regional heterogeneity result in microenvironmental selection and hence, emergence of genetic variations within tumors. The concept of adaptive response is an important departure from the traditional view that genetic heterogeneity is the product of increased random mutations, which implies that molecular heterogeneity is fundamentally unpredictable and, therefore, chaotic. The Darwinian model proposes that genetic heterogeneity is the result of a predictable and reproducible selection of successful adaptive strategies to local microenvironmental conditions.

Current cross-sectional imaging modalities can be used to identify regional variations in selection forces by using contrast-enhanced, cell density–based, or metabolic features. Clinical imaging can also be used to identify evidence of cellular adaptation. For example, if a region of low perfusion on a contrast-enhanced study is necrotic, then an adaptive population is absent or minimal. However, if the poorly perfused area is cellular, then there is presumptive evidence of an adapted proliferating population. While the specific genetic properties of this population cannot be determined, the phenotype of the adaptive strategy is predictable since the environmental conditions are more or less known. Thus, standard medical images can be used to infer specific emergent phenotypes and, with ongoing research, these phenotypes can be associated with underlying genetic changes.

This area of investigation will likely be challenging. As noted earlier, the most obvious spatially heterogeneous imaging feature in tumors is perfusion heterogeneity on contrast-enhanced CT or MR images. It generally has been assumed that the links between contrast enhancement, blood flow, perfusion, and tumor cell characteristics are straightforward. That is, tumor regions with decreased blood flow will exhibit low perfusion, low cell density, and high necrosis. In reality, however, the dynamics are actually much more complex. As shown in Figure 4, when using multiple superimposed sequences from MR imaging of malignant gliomas, regions of tumor that are poorly perfused on contrast-enhanced T1-weighted images may exhibit areas of low or high water content on T2-weighted images and low or high diffusion on diffusion-weighted images. Thus, high or low cell densities can coexist in poorly perfused volumes, creating perfusion-diffusion mismatches. Regions with poor perfusion with high cell density are of particular clinical interest because they represent a cell population that is apparently adapted to microenvironmental conditions associated with poor perfusion. The associated hypoxia, acidosis, and nutrient deprivation select for cells that are resistant to apoptosis and thus are likely to be resistant to therapy (46,47).


Figure 4: Left: Contrast-enhanced T1 image from subject TCGA-02-0034 in The Cancer Genome Atlas–Glioblastoma Multiforme repository of MR volumes of glioblastoma multiforme cases. Right: Spatial distribution of MR imaging–defined habitats within the tumor. The blue region (low T1 postgadolinium, low fluid-attenuated inversion recovery) is particularly notable because it presumably represents a habitat with low blood flow but high cell density, indicating a population presumably adapted to hypoxic acidic conditions.

Furthermore, other selection forces not related to perfusion are likely to be present within tumors. For example, evolutionary models suggest that cancer cells, even in stable microenvironments, tend to speciate into “engineers” that maximize tumor cell growth by promoting angiogenesis and “pioneers” that proliferate by invading normal issue and co-opting the blood supply. These invasive tumor phenotypes can exist only at the tumor edge, where movement into a normal tissue microenvironment can be rewarded by increased proliferation. This evolutionary dynamic may contribute to distinct differences between the tumor edges and the tumor cores, which frequently can be seen at analysis of cross-sectional images (Fig 5).


Figure 5a: CT images obtained with conventional entropy filtering in two patients with non–small cell lung cancer with no apparent textural differences show similar entropy values across all sections. 


Figure 5b: Contour plots obtained after the CT scans were convolved with the entropy filter. Further subdividing each section in the tumor stack into tumor edge and core regions (dotted black contour) reveals varying textural behavior across sections. Two distinct patterns have emerged, and preliminary analysis shows that the change of mean entropy value between core and edge regions correlates negatively with survival.

Interpretation of the subsegmentation of tumors will require computational models to understand and predict the complex nonlinear dynamics that lead to heterogeneous combinations of radiographic features. We have exploited ecologic methods and models to investigate regional variations in cancer environmental and cellular properties that lead to specific imaging characteristics. Conceptually, this approach assumes that regional variations in tumors can be viewed as a coalition of distinct ecologic communities or habitats of cells in which the environment is governed, at least to first order, by variations in vascular density and blood flow. The environmental conditions that result from alterations in blood flow, such as hypoxia, acidosis, immune response, growth factors, and glucose, represent evolutionary selection forces that give rise to local-regional phenotypic adaptations. Phenotypic alterations can result from epigenetic, genetic, or chromosomal rearrangements, and these in turn will affect prognosis and response to therapy. Changes in habitats or the relative abundance of specific ecologic communities over time and in response to therapy may be a valuable metric with which to measure treatment efficacy and emergence of resistant populations.


Emerging Strategies for Tumor Habitat Characterization

A method for converting images to spatially explicit tumor habitats is shown in Figure 4. Here, three-dimensional MR imaging data sets from a glioblastoma are segmented. Each voxel in the tumor is defined by a scale that includes its image intensity in different sequences. In this case, the imaging sets are from (a) a contrast-enhanced T1 sequence, (b) a fast spin-echo T2 sequence, and (c) a fluid-attenuated inversion-recovery (or FLAIR) sequence. Voxels in each sequence can be defined as high or low based on their value compared with the mean signal value. By using just two sequences, a contrast-enhanced T1 sequence and a fluid-attenuated inversion-recovery sequence, we can define four habitats: high or low postgadolinium T1 divided into high or low fluid-attenuated inversion recovery. When these voxel habitats are projected into the tumor volume, we find they cluster into spatially distinct regions. These habitats can be evaluated both in terms of their relative contributions to the total tumor volume and in terms of their interactions with each other, based on the imaging characteristics at the interfaces between regions. Similar spatially explicit analysis can be performed with CT scans (Fig 5).

Analysis of spatial patterns in cross-sectional images will ultimately require methods that bridge spatial scales from microns to millimeters. One possible method is a general class of numeric tools that is already widely used in terrestrial and marine ecology research to link species occurrence or abundance with environmental parameters. Species distribution models (4851) are used to gain ecologic and evolutionary insights and to predict distributions of species or morphs across landscapes, sometimes extrapolating in space and time. They can easily be used to link the environmental selection forces in MR imaging-defined habitats to the evolutionary dynamics of cancer cells.


Imaging can have an enormous role in the development and implementation of patient-specific therapies in cancer. The achievement of this goal will require new methods that expand and ultimately replace the current subjective qualitative assessments of tumor characteristics. The need for quantitative imaging has been clearly recognized by the National Cancer Institute and has resulted in formation of the Quantitative Imaging Network. A critical objective of this imaging consortium is to use objective, reproducible, and quantitative feature metrics extracted from clinical images to develop patient-specific imaging-based prognostic models and personalized cancer therapies.

It is increasingly clear that tumors are not homogeneous organlike systems. Rather, they contain regional coalitions of ecologic communities that consist of evolving cancer, stroma, and immune cell populations. The clinical consequence of such niche variations is that spatial and temporal variations of tumor phenotypes will inevitably evolve and present substantial challenges to targeted therapies. Hence, future research in cancer imaging will likely focus on spatially explicit analysis of tumor regions.

Clinical imaging can readily characterize regional variations in blood flow, cell density, and necrosis. When viewed in a Darwinian evolutionary context, these features reflect regional variations in environmental selection forces and can, at least in principle, be used to predict the likely adaptive strategies of the local cancer population. Hence, analyses of radiologic data can be used to inform evolutionary models and then can be mapped to regional population dynamics. Ecologic and evolutionary principles may provide a theoretical framework to link imaging to the cellular and molecular features of cancer cells and ultimately lead to a more comprehensive understanding of specific cancer biology in individual patients.



  • • Marked heterogeneity in genetic properties of different cells in the same tumor is typical and reflects ongoing intratumoral evolution.
  • • Evolution within tumors is governed by Darwinian dynamics, with identifiable environmental selection forces that interact with phenotypic (not genotypic) properties of tumor cells in a predictable and reproducible manner; clinical imaging is uniquely suited to measure temporal and spatial heterogeneity within tumors that is both a cause and a consequence of this evolution.
  • • Subjective radiologic descriptors of cancers are inadequate to capture this heterogeneity and must be replaced by quantitative metrics that enable statistical comparisons between features describing intratumoral heterogeneity and clinical outcomes and molecular properties.
  • • Spatially explicit mapping of tumor regions, for example by superimposing multiple imaging sequences, may permit patient-specific characterization of intratumoral evolution and ecology, leading to patient- and tumor-specific therapies.
  • • We summarize current information on quantitative analysis of radiologic images and propose future quantitative imaging must become spatially explicit to identify intratumoral habitats before and during therapy.

Disclosures of Conflicts of Interest: R.A.G. No relevant conflicts of interest to disclose. O.G. No relevant conflicts of interest to disclose.R.J.G. No relevant conflicts of interest to disclose.



The authors thank Mark Lloyd, MS; Joel Brown, PhD; Dmitry Goldgoff, PhD; and Larry Hall, PhD, for their input to image analysis and for their lively and informative discussions.


  • Received December 18, 2012; revision requested February 5, 2013; revision received March 11; accepted April 9; final version accepted April 29.
  • Funding: This research was supported by the National Institutes of Health (grants U54CA143970-01, U01CA143062; R01CA077575, andR01CA170595).


    1. Kurland BF,
    2. Gerstner ER,
    3. Mountz JM,
    4. et al

    . Promise and pitfalls of quantitative imaging in oncology clinical trials. Magn Reson Imaging2012;30(9):1301–1312.

    1. Levy MA,
    2. Freymann JB,
    3. Kirby JS,
    4. et al

    . Informatics methods to enable sharing of quantitative imaging research data. Magn Reson Imaging2012;30(9):1249–1256.

    1. Mirnezami R,
    2. Nicholson J,
    3. Darzi A

    . Preparing for precision medicine. N Engl J Med 2012;366(6):489–491.

    1. Yachida S,
    2. Jones S,
    3. Bozic I,
    4. et al

    . Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature 2010;467(7319):1114–1117.

    1. Gerlinger M,
    2. Rowan AJ,
    3. Horswell S,
    4. et al

    . Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med2012;366(10):883–892.

    1. Gerlinger M,
    2. Swanton C

    . How Darwinian models inform therapeutic failure initiated by clonal heterogeneity in cancer medicine. Br J Cancer2010;103(8):1139–1143.

    1. Kern SE

    . Why your new cancer biomarker may never work: recurrent patterns and remarkable diversity in biomarker failures. Cancer Res2012;72(23):6097–6101.

    1. Nowell PC

    . The clonal evolution of tumor cell populations. Science1976;194(4260):23–28.

    1. Greaves M,
    2. Maley CC

    . Clonal evolution in cancer. Nature2012;481(7381):306–313.

    1. Vincent TL,
    2. Brown JS

    . Evolutionary game theory, natural selection and Darwinian dynamics. Cambridge, England: Cambridge University Press, 2005.

    1. Gatenby RA,
    2. Gillies RJ

    . A microenvironmental model of carcinogenesis. Nat Rev Cancer 2008;8(1):56–61.

    1. Bowers MA,
    2. Matter SF

    . Landscape ecology of mammals: relationships between density and patch size. J Mammal 1997;78(4):999–1013.

    1. Dorner BK,
    2. Lertzman KP,
    3. Fall J

    . Landscape pattern in topographically complex landscapes: issues and techniques for analysis. Landscape Ecol2002;17(8):729–743.

    1. González-García I,
    2. Solé RV,
    3. Costa J

    . Metapopulation dynamics and spatial heterogeneity in cancer. Proc Natl Acad Sci U S A2002;99(20):13085–13089.

    1. Patel LR,
    2. Nykter M,
    3. Chen K,
    4. Zhang W

    . Cancer genome sequencing: understanding malignancy as a disease of the genome, its conformation, and its evolution. Cancer Lett 2012 Oct 27. [Epub ahead of print]

    1. Jaffe CC

    . Measures of response: RECIST, WHO, and new alternatives. J Clin Oncol 2006;24(20):3245–3251.

    1. Burton A

    . RECIST: right time to renovate? Lancet Oncol2007;8(6):464–465.

    1. Lambin P,
    2. Rios-Velazquez E,
    3. Leijenaar R,
    4. et al

    . Radiomics: extracting more information from medical images using advanced feature analysis. Eur J Cancer 2012;48(4):441–446.

    1. Nair VS,
    2. Gevaert O,
    3. Davidzon G,
    4. et al

    . Prognostic PET 18F-FDG uptake imaging features are associated with major oncogenomic alterations in patients with resected non-small cell lung cancer. Cancer Res2012;72(15):3725–3734.

    1. Diehn M,
    2. Nardini C,
    3. Wang DS,
    4. et al

    . Identification of noninvasive imaging surrogates for brain tumor gene-expression modules. Proc Natl Acad Sci U S A 2008;105(13):5213–5218.

    1. Segal E,
    2. Sirlin CB,
    3. Ooi C,
    4. et al

    . Decoding global gene expression programs in liver cancer by noninvasive imaging. Nat Biotechnol 2007;25(6):675–680.

    1. Tixier F,
    2. Le Rest CC,
    3. Hatt M,
    4. et al

    . Intratumor heterogeneity characterized by textural features on baseline 18F-FDG PET images predicts response to concomitant radiochemotherapy in esophageal cancer. J Nucl Med2011;52(3):369–378.

    1. Pang KK,
    2. Hughes T

    . MR imaging of the musculoskeletal soft tissue mass: is heterogeneity a sign of malignancy? J Chin Med Assoc2003;66(11):655–661.

    1. Ganeshan B,
    2. Panayiotou E,
    3. Burnand K,
    4. Dizdarevic S,
    5. Miles K

    . Tumour heterogeneity in non-small cell lung carcinoma assessed by CT texture analysis: a potential marker of survival. Eur Radiol 2012;22(4):796–802.

    1. Asselin MC,
    2. O’Connor JP,
    3. Boellaard R,
    4. Thacker NA,
    5. Jackson A

    . Quantifying heterogeneity in human tumours using MRI and PET. Eur J Cancer2012;48(4):447–455.

    1. Ahmed A,
    2. Gibbs P,
    3. Pickles M,
    4. Turnbull L

    . Texture analysis in assessment and prediction of chemotherapy response in breast cancer. J Magn Reson Imaging doi:10.1002/jmri.23971 2012. Published online December 13, 2012.

    1. Kawata Y,
    2. Niki N,
    3. Ohmatsu H,
    4. et al

    . Quantitative classification based on CT histogram analysis of non-small cell lung cancer: correlation with histopathological characteristics and recurrence-free survival. Med Phys2012;39(2):988–1000.

    1. Rubin DL

    . Creating and curating a terminology for radiology: ontology modeling and analysis. J Digit Imaging 2008;21(4):355–362.

    1. Opulencia P,
    2. Channin DS,
    3. Raicu DS,
    4. Furst JD

    . Mapping LIDC, RadLex™, and lung nodule image features. J Digit Imaging 2011;24(2):256–270.

    1. Channin DS,
    2. Mongkolwat P,
    3. Kleper V,
    4. Rubin DL

    . The Annotation and Image Mark-up project. Radiology 2009;253(3):590–592.

    1. Rubin DL,
    2. Mongkolwat P,
    3. Kleper V,
    4. Supekar K,
    5. Channin DS

    . Medical imaging on the semantic web: annotation and image markup. Presented at the AAAI Spring Symposium Series, Semantic Scientific Knowledge Integration, Palo Alto, Calif, March 26–28, 2008.

    1. Goh V,
    2. Ganeshan B,
    3. Nathan P,
    4. Juttla JK,
    5. Vinayan A,
    6. Miles KA

    . Assessment of response to tyrosine kinase inhibitors in metastatic renal cell cancer: CT texture as a predictive biomarker. Radiology 2011;261(1):165–171.

    1. Miles KA,
    2. Ganeshan B,
    3. Griffiths MR,
    4. Young RC,
    5. Chatwin CR

    . Colorectal cancer: texture analysis of portal phase hepatic CT images as a potential marker of survival. Radiology 2009;250(2):444–452.

    1. Haralick RM,
    2. Shanmugam K,
    3. Dinstein I

    . Textural features for image classification. IEEE Trans Syst Man Cybern 1973;3(6):610–621.

    1. Yang X,
    2. Knopp MV

    . Quantifying tumor vascular heterogeneity with dynamic contrast-enhanced magnetic resonance imaging: a review. J Biomed Biotechnol 2011;2011:732848.

    1. Frouin F,
    2. Bazin JP,
    3. Di Paola M,
    4. Jolivet O,
    5. Di Paola R

    . FAMIS: a software package for functional feature extraction from biomedical multidimensional images. Comput Med Imaging Graph 1992;16(2):81–91.

    1. Frouge C,
    2. Guinebretière JM,
    3. Contesso G,
    4. Di Paola R,
    5. Bléry M

    . Correlation between contrast enhancement in dynamic magnetic resonance imaging of the breast and tumor angiogenesis. Invest Radiol 1994;29(12):1043–1049.

    1. Zagdanski AM,
    2. Sigal R,
    3. Bosq J,
    4. Bazin JP,
    5. Vanel D,
    6. Di Paola R

    . Factor analysis of medical image sequences in MR of head and neck tumors. AJNR Am J Neuroradiol 1994;15(7):1359–1368.

    1. Bonnerot V,
    2. Charpentier A,
    3. Frouin F,
    4. Kalifa C,
    5. Vanel D,
    6. Di Paola R

    . Factor analysis of dynamic magnetic resonance imaging in predicting the response of osteosarcoma to chemotherapy. Invest Radiol 1992;27(10):847–855.

    1. Furman-Haran E,
    2. Grobgeld D,
    3. Kelcz F,
    4. Degani H

    . Critical role of spatial resolution in dynamic contrast-enhanced breast MRI. J Magn Reson Imaging2001;13(6):862–867.

    1. Rose CJ,
    2. Mills SJ,
    3. O’Connor JPB,
    4. et al

    . Quantifying spatial heterogeneity in dynamic contrast-enhanced MRI parameter maps. Magn Reson Med2009;62(2):488–499.

    1. Canuto HC,
    2. McLachlan C,
    3. Kettunen MI,
    4. et al

    . Characterization of image heterogeneity using 2D Minkowski functionals increases the sensitivity of detection of a targeted MRI contrast agent. Magn Reson Med2009;61(5):1218–1224.

    1. Lloyd MC,
    2. Allam-Nandyala P,
    3. Purohit CN,
    4. Burke N,
    5. Coppola D,
    6. Bui MM

    . Using image analysis as a tool for assessment of prognostic and predictive biomarkers for breast cancer: how reliable is it? J Pathol Inform2010;1:29–36.

    1. Kumar V,
    2. Nath K,
    3. Berman CG,
    4. et al

    . Variance of SUVs for FDG-PET/CT is greater in clinical practice than under ideal study settings. Clin Nucl Med2013;38(3):175–182.

    1. Walker-Samuel S,
    2. Orton M,
    3. Boult JK,
    4. Robinson SP

    . Improving apparent diffusion coefficient estimates and elucidating tumor heterogeneity using Bayesian adaptive smoothing. Magn Reson Med 2011;65(2):438–447.

    1. Thews O,
    2. Nowak M,
    3. Sauvant C,
    4. Gekle M

    . Hypoxia-induced extracellular acidosis increases p-glycoprotein activity and chemoresistance in tumors in vivo via p38 signaling pathway. Adv Exp Med Biol 2011;701:115–122.

    1. Thews O,
    2. Dillenburg W,
    3. Rösch F,
    4. Fellner M

    . PET imaging of the impact of extracellular pH and MAP kinases on the p-glycoprotein (Pgp) activity. Adv Exp Med Biol 2013;765:279–286.

    1. Araújo MB,
    2. Peterson AT

    . Uses and misuses of bioclimatic envelope modeling. Ecology 2012;93(7):1527–1539.

    1. Larsen PE,
    2. Gibbons SM,
    3. Gilbert JA

    . Modeling microbial community structure and functional diversity across time and space. FEMS Microbiol Lett2012;332(2):91–98.

    1. Shenton W,
    2. Bond NR,
    3. Yen JD,
    4. Mac Nally R

    . Putting the “ecology” into environmental flows: ecological dynamics and demographic modelling. Environ Manage 2012;50(1):1–10.

    1. Clark MC,
    2. Hall LO,
    3. Goldgof DB,
    4. Velthuizen R,
    5. Murtagh FR,
    6. Silbiger MS

    .Automatic tumor segmentation using knowledge-based techniques. IEEE Trans Med Imaging 1998;17(2):187–201.

Read Full Post »

Follow-up on Tomosynthesis

Writer & Curator: Dror Nir, PhD

Tomosynthesis, is a method for performing high-resolution limited-angle (i.e. not full 3600 rotation but more like ~500) tomography. The use of such systems in breast-cancer screening is steadily increasing following the clearance of such system by the FDA on 2011; see my posts – Improving Mammography-based imaging for better treatment planning and State of the art in oncologic imaging of breast.

Many radiologists expects that Tomosynthesis will eventually replace conventional mammography due to the fact that it increases the sensitivity of breast cancer detection. This claim is supported by new peer-reviewed publications. In addition, the patient’s experience during Tomosynthesis is less painful due to a lesser pressure that is applied to the breast and while presented with higher in-plane resolution and less imaging artifacts the mean glandular dose of digital breast Tomosynthesis is comparable to that of full field digital mammography. Because it is relatively new, Tomosynthesis is not available at every hospital. As well, the procedure is recognized for reimbursement by public-health schemes.

A good summary of radiologist opinion on Tomosynthesis can be found in the following video:

Recent studies’ results with digital Tomosynthesis are promising. In addition to increase in sensitivity for detection of small cancer lesions researchers claim that this new breast imaging technique will make breast cancers easier to see in dense breast tissue.  Here is a paper published on-line by the Lancet just a couple of months ago:

Integration of 3D digital mammography with tomosynthesis for population breast-cancer screening (STORM): a prospective comparison study

Stefano Ciatto†, Nehmat Houssami, Daniela Bernardi, Francesca Caumo, Marco Pellegrini, Silvia Brunelli, Paola Tuttobene, Paola Bricolo, Carmine Fantò, Marvi Valentini, Stefania Montemezzi, Petra Macaskill , Lancet Oncol. 2013 Jun;14(7):583-9. doi: 10.1016/S1470-2045(13)70134-7. Epub 2013 Apr 25.

Background Digital breast tomosynthesis with 3D images might overcome some of the limitations of conventional 2D mammography for detection of breast cancer. We investigated the effect of integrated 2D and 3D mammography in population breast-cancer screening.

Methods Screening with Tomosynthesis OR standard Mammography (STORM) was a prospective comparative study. We recruited asymptomatic women aged 48 years or older who attended population-based breast-cancer screening through the Trento and Verona screening services (Italy) from August, 2011, to June, 2012. We did screen-reading in two sequential phases—2D only and integrated 2D and 3D mammography—yielding paired data for each screen. Standard double-reading by breast radiologists determined whether to recall the participant based on positive mammography at either screen read. Outcomes were measured from final assessment or excision histology. Primary outcome measures were the number of detected cancers, the number of detected cancers per 1000 screens, the number and proportion of false positive recalls, and incremental cancer detection attributable to integrated 2D and 3D mammography. We compared paired binary data with McNemar’s test.

Findings 7292 women were screened (median age 58 years [IQR 54–63]). We detected 59 breast cancers (including 52 invasive cancers) in 57 women. Both 2D and integrated 2D and 3D screening detected 39 cancers. We detected 20 cancers with integrated 2D and 3D only versus none with 2D screening only (p<0.0001). Cancer detection rates were 5·3 cancers per 1000 screens (95% CI 3.8–7.3) for 2D only, and 8.1 cancers per 1000 screens (6.2–10.4) for integrated 2D and 3D screening. The incremental cancer detection rate attributable to integrated 2D and 3D mammography was 2.7 cancers per 1000 screens (1.7–4.2). 395 screens (5.5%; 95% CI 5.0–6.0) resulted in false positive recalls: 181 at both screen reads, and 141 with 2D only versus 73 with integrated 2D and 3D screening (p<0·0001). We estimated that conditional recall (positive integrated 2D and 3D mammography as a condition to recall) could have reduced false positive recalls by 17.2% (95% CI 13.6–21.3) without missing any of the cancers detected in the study population.

Interpretation Integrated 2D and 3D mammography improves breast-cancer detection and has the potential to reduce false positive recalls. Randomised controlled trials are needed to compare integrated 2D and 3D mammography with 2D mammography for breast cancer screening.

Funding National Breast Cancer Foundation, Australia; National Health and Medical Research Council, Australia; Hologic, USA; Technologic, Italy.


Although controversial, mammography screening is the only population-level early detection strategy that has been shown to reduce breast-cancer mortality in randomised trials.1,2 Irrespective of which side of the mammography screening debate one supports,1–3 efforts should be made to investigate methods that enhance the quality of (and hence potential benefit from) mam­mography screening. A limitation of standard 2D mammography is the superimposition of breast tissue or parenchymal density, which can obscure cancers or make normal structures appear suspicious. This short coming reduces the sensitivity of mammography and increases false-positive screening. Digital breast tomosynthesis with 3D images might help to overcome these limitations. Several reviews4,5 have described the development of breast tomosynthesis technology, in which several low-dose radiographs are used to reconstruct a pseudo-3D image of the breast.4–6

Initial clinical studies of 3D mammography, 6–10 though based on small or selected series, suggest that addition of 3D to 2D mammography could improve cancer detection and reduce the number of false positives. However, previous assessments of breast tomosynthesis might have been constrained by selection biases that distorted the potential effect of 3D mammography; thus, screening trials of integrated 2D and 3D mammography are needed.6

We report the results of a large prospective study (Screening with Tomosynthesis OR standard Mammog­raphy [STORM]) of 3D digital mammography. We investi­gated the effect of screen-reading using both standard 2D and 3D imaging with tomosynthesis compared with screening with standard 2D digital mammography only for population breast-cancer screening.



Study design and participants

STORM is a prospective population-screening study that compares mammography screen-reading in two sequential phases (figure)—2D only versus integrated 2D and 3D mammography with tomosynthesis—yielding paired results for each screening examination. Women aged 48 years or older who attended population-based screening through the Trento and Verona screening services, Italy, from August, 2011, to June, 2012, were invited to be screened with integrated 2D and 3D mammography. Participants in routine screening mammography (once every 2 years) were asymptomatic women at standard (population) risk for breast cancer. The study was granted institutional ethics approval at each centre, and participants gave written informed consent. Women who opted not to participate in the study received standard 2D mammography. Digital mammography has been used in the Trento breast-screening programme since 2005, and in the Verona programme since 2007; each service monitors outcomes and quality indicators as dictated by European standards, and both have published data for screening performance.11,12


study design


All participants had digital mammography using a Selenia Dimensions Unit with integrated 2D and 3D mammography done in the COMBO mode (Hologic, Bedford, MA, USA): this setting takes 2D and 3D images at the same screening examination with a single breast position and compression. Each 2D and 3D image consisted of a bilateral two-view (mediolateral oblique and craniocaudal) mammogram. Screening mammo­grams were interpreted sequentially by radiologists, first on the basis of standard 2D mammography alone, and then by the same radiologist (on the same day) on the basis of integrated 2D and 3D mammography (figure). Thus, integrated 2D and 3D mammography screening refers to non-independent screen reading based on joint interpretation of 2D and 3D images, and does not refer to analytical combinations. Radiologists had to record whether or not to recall the participant at each screen-reading phase before progressing to the next phase of the sequence. For each screen, data were also collected for breast density (at the 2D screen-read), and the side and quadrant for any recalled abnormality (at each screen-read). All eight radiologists were breast radiologists with a mean of 8 years (range 3–13 years) experience in mammography screening, and had received basic training in integrated 2D and 3D mammography. Several of the radiologists had also used 2D and 3D mammography for patients recalled after positive conventional mammography screening as part of previous studies of tomosynthesis.8,13

Mammograms were interpreted in two independent screen-reads done in parallel, as practiced in most population breast-screening programs in Europe. A screen was considered positive and the woman recalled for further investigations if either screen-reader recorded a positive result at either 2D or integrated 2D and 3D screening (figure). When previous screening mammograms were available, these were shown to the radiologist at the time of screen-reading, as is standard practice. For assessment of breast density, we used Breast Imaging Reporting and Data System (BI-RADS)14 classification, with participants allocated to one of two groups (1–2 [low density] or 3–4 [high density]). Disagreement between readers about breast density was resolved by assessment by a third reader.

Our primary outcomes were the number of cancers detected, the number of cancers detected per 1000 screens, the number and percentage of false posi­tive recalls, and the incremental cancer detection rate attributable to integrated 2D and 3D mammography screening. We compared the number of cancers that were detected only at 2D mammography screen-reading and those that were detected only at 2D and 3D mammography screen-reading; we also did this analysis for false positive recalls. To explore the potential effect of integrated 2D and 3D screening on false-positive recalls, we also estimated how many false-positive recalls would have resulted from using a hypothetical conditional false-positive recall approach; – i.e. positive integrated 2D and 3D mammography as a condition of recall (screening recalled at 2D mammography only would not be recalled). Pre-planned secondary analyses were comparison of outcome measures by age group and breast density.

Outcomes were assessed by excision histology for participants who had surgery, or the complete assessment outcome (including investigative imaging with or without histology from core needle biopsy) for all recalled participants. Because our study focuses on the difference in detection by the two screening methods, some cancers might have been missed by both 2D and integrated 2D and 3D mammography; this possibility could be assessed at future follow-up to identify interval cancers. However, this outcome is not assessed in the present study and does not affect estimates of our primary outcomes – i.e. comparative true or false positive detection for 2D-only versus integrated 2D and 3D mammography.


Statistical analysis

The sample size was chosen to provide 80% power to detect a difference of 20% in cancer detection, assuming a detection probability of 80% for integrated 2D and 3D screening mammography and 60% for 2D only screening, with a two-sided significance threshold of 5%. Based on the method of Lachenbruch15 for estimating sample size for studies that use McNemar’s test for paired binary data, a minimum of 40 cancers were needed. Because most screens in the participating centres were incident (repeat) screening (75%–80%), we used an underlying breast-cancer prevalence of 0·5% to estimate that roughly 7500–8000 screens would be needed to identify 40 cancers in the study population.

We calculated the Wilson CI for the false-positive recall ratio for integrated 2D and 3D screening with conditional recall compared with 2D only screening.16 All of the other analyses were done with SAS/STAT (version 9.2), using exact methods to compute 95 CIs and p-values.

Role of the funding source

The sponsors of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author (NH) had full access to all the data in the study and had final responsibility for the decision to submit for publication.


7292 participants with a median age of 58 years (IQR 54–63, range 48–71) were screened between Aug 12, 2011, and June 29, 2012. Roughly 5% of invited women declined integrated 2D and 3D screening and received standard 2D mammography. We present data for 7294 screens because two participants had bilateral cancer (detected with different screen-reading techniques for one participant). We detected 59 breast cancers in 57 participants (52 invasive cancers and seven ductal carcinoma in-situ). Of the invasive cancers, most were invasive ductal (n=37); others were invasive special types (n=7), invasive lobular (n=4), and mixed invasive types (n=4).

Table 1 shows the characteristics of the cancers. Mean tumour size (for the invasive cancers with known exact size) was 13.7 mm (SD 5.8) for cancers detected with both 2D alone and integrated 2D and 3D screening (n=29), and 13.5 mm (SD 6.7) for cancers detected only with integrated 2D and 3D screening (n=13).


Table 1

Of the 59 cancers, 39 were detected at both 2D and integrated 2D and 3D screening (table 2). 20 cancers were detected with only integrated 2D and 3D screening compared with none detected with only 2D screening (p<0.0001; table 2). 395 screens were false positive (5.5%, 95% CI 5.0–6.0); 181 occurred at both screen-readings, and 141 occurred at 2D screening only compared with 73 at integrated 2D and 3D screening (p<0.0001; table 2). These differences were still significant in sensitivity analyses that excluded the two participants with bilateral cancer (data not shown).

Table 2

5.3 cancers per 1000 screens (95% CI 3.8–7.3; table 3) were detected with 2D mammography only versus 8.1 cancers per 1000 screens (95% CI 6.2–10.4) with integrated 2D and 3D mammography (p<0.0001). The incremental cancer detection rate attributable to inte­grated 2D and 3D screening was 2.7 cancers per 1000 screens (95% CI 1.7–4.2), which is 33.9% (95% CI 22.1–47.4) of the cancers detected in the study popu­lation. In a sensitivity analysis that excluded the two participants with bilateral cancer the estimated incre­mental cancer detection rate attributable to integrated 2D and 3D screening was 2.6 cancers per 1000 screens (95% CI 1.4–3.8). The stratified results show that integrated 2D and 3D mammography was associated with an incrementally increased cancer detection rate in both age-groups and density categories (tables 3–5). A minority (16.7%) of breasts were of high density (category 3–4) reducing the power of statistical comparisons in this subgroup (table 5). The incremental cancer detection rate was much the same in low density versus high density groups (2.8 per 1000 vs 2.5 per 1000; p=0.84; table 3).

Table 3

Table 4-5

Overall recall—any recall resulting in true or false positive screens—was 6.2% (95% CI 5.7–6.8), and the false-positive rate for the 7235 screens of participants who did not have breast cancer was 5.5% (5.0–6.0). Table 6 shows the contribution to false-positive recalls from 2D mammography only, integrated 2D and 3D mammography only, and both, and the estimated number of false positives if positive integrated 2D and 3D mammography was a condition for recall (positive 2D only not recalled). Overall, more of the false-positive rate was driven by 2D mammography only than by integrated 2D and 3D, although almost half of the false-positive rate was a result of false positives recalled at both screen-reading phases (table 6). The findings were much the same when stratified by age and breast density (table 6). Had a conditional recall rule been applied, we estimate that the false-positive rate would have been 3.5% (95% CI 3.1–4.0%; table 6) and could have potentially prevented 68 of the 395 false positives (a reduction of 17.2%; 95% CI 13.6–21.3). The ratio between the number of false positives with integrated 2D and 3D screening with conditional recall (n=254) versus 2D only screening (n=322) was 0.79 (95% CI 0.71–0.87).


Our study showed that integrated 2D and 3D mam­mography screening significantly increases detection of breast cancer compared with conventional mammog­raphy screening. There was consistent evidence of an incremental improvement in detection from integrated 2D and 3D mammography across age-group and breast density strata, although the analysis by breast density was limited by low number of women with breasts of high density.

One should note that we investigated comparative cancer detection, and not absolute screening sensitivity. By integrating 2D and 3D mammography using the study screen-reading protocol, 1% of false-positive recalls resulted from 2D and 3D screen-reading only (table 6). However, significantly more false positives resulted from 2D only mammography compared with integrated 2D and 3D mammography, both overall and in the stratified analyses. Application of a conditional recall rule would have resulted in a false-positive rate of 3.5% instead of the actual false-positive rate of 5.5%. The estimated false positive recall ratio of 0.79 for integrated 2D and 3D screening with conditional recall compared with 2D only screening suggests that integrated 2D and 3D screening could reduce false recalls by roughly a fifth. Had such a condition been adopted, none of the cancers detected in the study would have been missed because no cancers were detected by 2D mammography only, although this result might be because our design allowed an independent read for 2D only mammography whereas the integrated 2D and 3D read was an interpretation of a combination of 2D and 3D imaging. We do not recommend that such a conditional recall rule be used in breast-cancer screening until our findings are replicated in other mammography screening studies—STORM involved double-reading by experienced breast radiologists, and our results might not apply to other screening settings. Using a test set of 130 mammograms, Wallis and colleagues7 report that adding tomosynthesis to 2D mammography increased the accuracy of inexperienced readers (but not of experienced readers), therefore having experienced radiologists in STORM could have underestimated the effect of integrated 2D and 3D screen-reading.

No other population screening trials of integrated 2D and 3D mammography have reported final results (panel); however, an interim analysis of the Oslo trial17 a large population screening study has shown that integrated 2D and 3D mammography substantially increases detection of breast cancer. The Oslo study investigators screened women with both 2D and 3D mammography, but randomised reading strategies (with vs without 3D mammograms) and adjusted for the different screen-readers,17whereas we used sequential screen-reading to keep the same reader for each exam­ination. Our estimates for comparative cancer detection and for cancer detection rates are consistent with those of the interim analysis of the Oslo study.17 The applied recall methods differed between the Oslo study (which used an arbitration meeting to decide recall) and the STORM study (we recalled based on a decision by either screen-reader), yet both studies show that 3D mammog­raphy reduces false-positive recalls when added to standard mammography.

An editorial in The Lancet18 might indeed signal the closing of a chapter of debate about the benefits and harms of screening. We hope that our work might be the beginning of a new chapter for mammography screening: our findings should encourage new assessments of screening using 2D and 3D mammography and should factor several issues related to our study. First, we compared standard 2D mammography with integrated 2D and 3D mammography the 3D mammograms were not interpreted independently of the 2D mammograms therefore 3D mammography only (without the 2D images) might not provide the same results. Our experience with breast tomosynthesis and a review6 of 3D mammography underscore the importance of 2D images in integrated 2D and 3D screen-reading. The 2D images form the basis of the radiologist’s ability to integrate the information from 3D images with that from 2D images. Second, although most screening in STORM was incident screening, the substantial increase in cancer detection rate with integrated 2D and 3D mammography results from the enhanced sensitivity of integrated 2D and 3D screening and is probably also a result of a prevalence effect (ie, the effect of a first screening round with integrated 2D and 3D mammography). We did not assess the effect of repeat (incident) screening with integrated 2D and 3D mammography on cancer detection it might provide a smaller effect on cancer detection rates than what we report. Third, STORM was not designed to measure biological differences between the cancers detected at integrated 2D and 3D screening compared with those detected at both screen-reading phases. Descriptive analyses suggest that, generally, breast cancers detected only at integrated 2D and 3D screening had similar features (eg, histology, pathological tumour size, node status) as those detected at both screen-reading phases. Thus, some of the cancers detected only at 2D and 3D screening might represent early detection (and would be expected to receive screening benefit) whereas some might represent over-detection and a harm from screening, as for conventional screening mam mography.1,19 The absence of consensus about over-diagnosis in breast-cancer screening should not detract from the importance of our study findings to applied screening research and to screening practice; however, our trial was not done to assess the extent to which integrated 2D and 3D mam­mography might contribute to over-diagnosis.

The average dose of glandular radiation from the many low-dose projections taken during a single acquisition of 3D mammography is roughly the same as that from 2D mammography.6,20–22 Using integrated 2D and 3D en­tails both a 2D and 3D acquisition in one breast com­pression, which roughly doubles the radiation dose to the breast. Therefore, integrated 2D and 3D mammography for population screening might only be justifiable if improved outcomes were not defined solely in terms of improved detection. For example, it would be valuable to show that the increased detection with integrated 2D and 3D screening leads to reduced interval cancer rates at follow-up. A limitation of our study might be that data for interval cancers were not available; however, because of the paired design we used, future evaluation of interval cancer rates from our study will only apply to breast cancers that were not identified using 2D only or integrated 2D and 3D screening. We know of two patients from our study who have developed interval cancers (follow-up range 8–16 months). We did not get this information from cancer registries and follow-up was very short, so these data should be interpreted very cautiously, especially because interval cancers would be expected to occur in the second year of the standard 2 year interval between screening rounds. Studies of interval cancer rates after integrated 2D and 3D mammography would need to be randomised controlled trials and have a very large sample size. Additionally, the development of reconstructed 2D images from a 3D mammogram23 provides a timely solution to concerns about radiation by providing both the 2D and 3D images from tomosynthesis, eliminating the need for two acquisitions.

We have shown that integrated 2D and 3D mammog­raphy in population breast-cancer screening increases detection of breast cancer and can reduce false-positive recalls depending on the recall strategy. Our results do not warrant an immediate change to breast-screening practice, instead, they show the urgent need for random­ised controlled trials of integrated 2D and 3D versus 2D mammography, and for further translational research in breast tomosynthesis. We envisage that future screening trials investigating this issue will include measures of breast cancer detection, and will be designed to assess interval cancer rates as a surrogate endpoint for screening efficacy.


SC had the idea for and designed the study, and collected and interpreted data. NH advised on study concepts and methods, analysed and interpreted data, searched the published work, and wrote and revised the report. DB and FC were lead radiologists, recruited participants, collected data, and commented on the draft report. MP, SB, PT, PB, PT, CF, and MV did the screen-reading, collected data, and reviewed the draft report. SM collected data and reviewed the draft report. PM planned the statistical analysis, analysed and interpreted data, and wrote and revised the report.

Conflicts of interest

SC, DB, FC, MP, SB, PT, PB, CF, MV, and SM received assistance from Hologic (Hologic USA; Technologic Italy) in the form of tomosynthesis technology and technical support for the duration of the study, and travel support to attend collaborators’ meetings. NH receives research support from a National Breast Cancer Foundation (NBCF Australia) Practitioner Fellowship, and has received travel support from Hologic to attend a collaborators’ meeting. PM receives research support through Australia’s National Health and Medical Research Council programme grant 633003 to the Screening & Test Evaluation Program.



1       Independent UK Panel on Breast Cancer Screening. The benefits and harms of breast cancer screening: an independent review. Lancet 2012; 380: 1778–86.

2       Glasziou P, Houssami N. The evidence base for breast cancer screening. Prev Med 2011; 53: 100–102.

3       Autier P, Esserman LJ, Flowers CI, Houssami N. Breast cancer screening: the questions answered. Nat Rev Clin Oncol 2012; 9: 599–605.

4       Baker JA, Lo JY. Breast tomosynthesis: state-of-the-art and review of the literature. Acad Radiol 2011; 18: 1298–310.

5       Helvie MA. Digital mammography imaging: breast tomosynthesis and advanced applications. Radiol Clin North Am 2010; 48: 917–29.

6      Houssami N, Skaane P. Overview of the evidence on digital breast tomosynthesis in breast cancer detection. Breast 2013; 22: 101–08.

7   Wallis MG, Moa E, Zanca F, Leifland K, Danielsson M. Two-view and single-view tomosynthesis versus full-field digital mammography: high-resolution X-ray imaging observer study. Radiology 2012; 262: 788–96.

8   Bernardi D, Ciatto S, Pellegrini M, et al. Prospective study of breast tomosynthesis as a triage to assessment in screening. Breast Cancer Res Treat 2012; 133: 267–71.

9   Michell MJ, Iqbal A, Wasan RK, et al. A comparison of the accuracy of film-screen mammography, full-field digital mammography, and digital breast tomosynthesis. Clin Radiol 2012; 67: 976–81.

10 Skaane P, Gullien R, Bjorndal H, et al. Digital breast tomosynthesis (DBT): initial experience in a clinical setting. Acta Radiol 2012; 53: 524–29.

11 Pellegrini M, Bernardi D, Di MS, et al. Analysis of proportional incidence and review of interval cancer cases observed within the mammography screening programme in Trento province, Italy. Radiol Med 2011; 116: 1217–25.

12 Caumo F, Vecchiato F, Pellegrini M, Vettorazzi M, Ciatto S, Montemezzi S. Analysis of interval cancers observed in an Italian mammography screening programme (2000–2006). Radiol Med 2009; 114: 907–14.

13 Bernardi D, Ciatto S, Pellegrini M, et al. Application of breast tomosynthesis in screening: incremental effect on mammography acquisition and reading time. Br J Radiol 2012; 85: e1174–78.

14 American College of Radiology. ACR BI-RADS: breast imaging reporting and data system, Breast Imaging Atlas. Reston: American College of Radiology, 2003.

15  Lachenbruch PA. On the sample size for studies based on McNemar’s test. Stat Med 1992; 11: 1521–25.

16  Bonett DG, Price RM. Confidence intervals for a ratio of binomial proportions based on paired data. Stat Med 2006; 25: 3039–47.

17  Skaane P, Bandos AI, Gullien R, et al. Comparison of digital mammography alone and digital mammography plus tomosynthesis in a population-based screening program. Radiology 2013; published online Jan 3. radiol.12121373.

18  The Lancet. The breast cancer screening debate: closing a chapter? Lancet 2012; 380: 1714.

19  Biesheuvel C, Barratt A, Howard K, Houssami N, Irwig L. Effects of study methods and biases on estimates of invasive breast cancer overdetection with mammography screening: a systematic review. Lancet Oncol 2007; 8: 1129–38.

20  Tagliafico A, Astengo D, Cavagnetto F, et al. One-to-one comparison between digital spot compression view and digital breast tomosynthesis. Eur Radiol 2012; 22: 539–44.

21  Tingberg A, Fornvik D, Mattsson S, Svahn T, Timberg P, Zackrisson S. Breast cancer screening with tomosynthesis—initial experiences. Radiat Prot Dosimetry 2011; 147: 180–83.

22  Feng SS, Sechopoulos I. Clinical digital breast tomosynthesis system: dosimetric characterization. Radiology 2012; 263: 35–42.

23  Gur D, Zuley ML, Anello MI, et al. Dose reduction in digital breast tomosynthesis (DBT) screening using synthetically reconstructed projection images: an observer performance study. Acad Radiol 2012; 19: 166–71.

A very good and down-to-earth comment on this article was made by Jules H Sumkin who disclosed that he is an unpaid member of SAB Hologic Inc and have a PI research agreement between University of Pittsburgh and Hologic Inc.

The results of the study by Stefano Ciatto and colleagues1 are consistent with recently published prospective,2,3 retrospective,4 and observational5 reports on the same topic. The study1 had limitations, including the fact that the same radiologist interpreted screens sequentially the same day without cross-balancing which examination was read first. Also, the false-negative findings for integrated 2D and 3D mammography, and therefore absolute benefit from the procedure, could not be adequately assessed because cases recalled by 2D mammography alone (141 cases) did not result in a single detection of an additional cancer while the recalls from the integrated 2D and 3D mammography alone (73 cases) resulted in the detection of 20 additional cancers. Nevertheless, the results are in strong agreement with other studies reporting of substantial performance improvements when the screening is done with integrated 2D and 3D mammography.

I disagree with the conclusion of the study with regards to the urgent need for randomised clinical trials of integrated 2D and 3D versus 2D mammography. First, to assess differences in mortality as a result of an imaging-based diagnostic method, a randomised trial will require several repeated screens by the same method in each study group, and the strong results from all studies to date will probably result in substantial crossover and self-selection biases over time. Second, because of the high survival rate (or low mortality rate) of breast cancer, the study will require long follow-up times of at least 10 years. In a rapidly changing environment in terms of improvements in screening technologies and therapeutic inter­ventions, the avoidance of biases is likely to be very difficult, if not impossible. The use of the number of interval cancers and possible shifts in stage at detection, while appropriately accounting for confounders, would be almost as daunting a task. Third, the imaging detection of cancer is only the first step in many management decisions and interventions that can affect outcome. The appropriate control of biases related to patient management is highly unlikely. The arguments above, in addition to the existing reports to date that show substantial improvements in cancer detection, particularly with the detection of invasive cancers, with a simultaneous reduction in recall rates, support the argument that a randomised trial is neither necessary nor warranted. The current technology might be obsolete by the time results of an appropriately done and analysed randomised trial is made public.

In order to better link the information given by “scientific” papers to the context of daily patients’ reality I suggest to spend some time reviewing few of the videos in the below links:

  1. The following group of videos is featured on a website by Siemens. Nevertheless, the presenting radiologists are leading practitioners who affects thousands of lives every year – What the experts say about tomosynthesis. – click on ECR 2013
  2. Breast Tomosynthesis in Practice – part of a commercial ad of the Washington Radiology Associates featured on the website of Diagnostic Imaging. As well, affects thousands of lives in the Washington area every year.

The pivotal questions yet to be answered are:

  1. What should be done in order to translate increase in sensitivity and early detection into decrease in mortality?

  2. What is the price of such increase in sensitivity in terms of quality of life and health-care costs and is it worth-while to pay?

An article that summarises positively the experience of introducing Tomosynthesis into routine screening practice was recently published on AJR:

Implementation of Breast Tomosynthesis in a Routine Screening Practice: An Observational Study

Stephen L. Rose1, Andra L. Tidwell1, Louis J. Bujnoch1, Anne C. Kushwaha1, Amy S. Nordmann1 and Russell Sexton, Jr.1

Affiliation: 1 All authors: TOPS Comprehensive Breast Center, 17030 Red Oak Dr, Houston, TX 77090.

Citation: American Journal of Roentgenology. 2013;200:1401-1408



OBJECTIVE. Digital mammography combined with tomosynthesis is gaining clinical acceptance, but data are limited that show its impact in the clinical environment. We assessed the changes in performance measures, if any, after the introduction of tomosynthesis systems into our clinical practice.

MATERIALS AND METHODS. In this observational study, we used verified practice- and outcome-related databases to compute and compare recall rates, biopsy rates, cancer detection rates, and positive predictive values for six radiologists who interpreted screening mammography studies without (n = 13,856) and with (n = 9499) the use of tomosynthesis. Two-sided analyses (significance declared at p < 0.05) accounting for reader variability, age of participants, and whether the examination in question was a baseline were performed.

RESULTS. For the group as a whole, the introduction and routine use of tomosynthesis resulted in significant observed changes in recall rates from 8.7% to 5.5% (p < 0.001), nonsignificant changes in biopsy rates from 15.2 to 13.5 per 1000 screenings (p = 0.59), and cancer detection rates from 4.0 to 5.4 per 1000 screenings (p = 0.18). The invasive cancer detection rate increased from 2.8 to 4.3 per 1000 screening examinations (p = 0.07). The positive predictive value for recalls increased from 4.7% to 10.1% (p < 0.001).

CONCLUSION. The introduction of breast tomosynthesis into our practice was associated with a significant reduction in recall rates and a simultaneous increase in breast cancer detection rates.

Here are the facts in tables and pictures from this article

Table 1 AJR

Table 2-3 AJR


Table 4 AJR


p1 ajr

p2 ajr

Other articles related to the management of breast cancer were published on this Open Access Online Scientific Journal:

Automated Breast Ultrasound System (‘ABUS’) for full breast scanning: The beginning of structuring a solution for an acute need!

Introducing smart-imaging into radiologists’ daily practice.

Not applying evidence-based medicine drives up the costs of screening for breast-cancer in the USA.

New Imaging device bears a promise for better quality control of breast-cancer lumpectomies – considering the cost impact

Harnessing Personalized Medicine for Cancer Management, Prospects of Prevention and Cure: Opinions of Cancer Scientific Leaders @

Predicting Tumor Response, Progression, and Time to Recurrence

“The Molecular pathology of Breast Cancer Progression”

Personalized medicine gearing up to tackle cancer

What could transform an underdog into a winner?

Mechanism involved in Breast Cancer Cell Growth: Function in Early Detection & Treatment

Nanotech Therapy for Breast Cancer

A Strategy to Handle the Most Aggressive Breast Cancer: Triple-negative Tumors

Breakthrough Technique Images Breast Tumors in 3-D With Great Clarity, Reduced Radiation

Closing the Mammography gap

Imaging: seeing or imagining? (Part 1)

Imaging: seeing or imagining? (Part 2)

Read Full Post »