Can Occult Invasive Disease in Ductal Carcinoma In Situ Be Predicted Using Computer-extracted Mammographic Features?

Application of deep learning on mammographies to discriminate between low and high-risk DCIS for patient participation in active surveillance trials

BACKGROUND

Ductal Carcinoma In Situ (DCIS) can progress to invasive breast cancer, but most DCIS lesions never will. Therefore, four clinical trials (COMET, LORIS, LORETTA, AND LORD) test whether active surveillance for women with low-risk Ductal carcinoma In Situ is safe (E. S. Hwang et al., BMJ Open, 9: e026797, 2019, A. Francis et al., Eur J Cancer. 51: 2296-2303, 2015, Chizuko Kanbayashi et al. The international collaboration of active surveillance trials for low-risk DCIS (LORIS, LORD, COMET, LORETTA),  L. E. Elshof et al., Eur J Cancer, 51, 1497-510, 2015). Low-risk is defined as grade I or II DCIS. Because DCIS grade is a major eligibility criteria in these trials, it would be very helpful to assess DCIS grade on mammography, informed by grade assessed on DCIS histopathology in pre-surgery biopsies, since surgery will not be performed on a significant number of patients participating in these trials.

OBJECTIVE

To assess the performance and clinical utility of a convolutional neural network (CNN) in discriminating high-risk (grade III) DCIS and/or Invasive Breast Cancer (IBC) from low-risk (grade I/II) DCIS based on mammographic features. We explored whether the CNN could be used as a decision support tool, from excluding high-risk patients for active surveillance.

METHODS

In this single centre retrospective study, 464 patients diagnosed with DCIS based on pre-surgery biopsy between 2000 and 2014 were included. The collection of mammography images was partitioned on a patient-level into two subsets, one for training containing 80% of cases (371 cases, 681 images) and 20% (93 cases, 173 images) for testing. A deep learning model based on the U-Net CNN was trained and validated on 681 two-dimensional mammograms. Classification performance was assessed with the Area Under the Curve (AUC) receiver operating characteristic and predictive values on the test set for predicting high risk DCIS-and high-risk DCIS and/ or IBC from low-risk DCIS.

RESULTS

When classifying DCIS as high-risk, the deep learning network achieved a Positive Predictive Value (PPV) of 0.40, Negative Predictive Value (NPV) of 0.91 and an AUC of 0.72 on the test dataset. For distinguishing high-risk and/or upstaged DCIS (occult invasive breast cancer) from low-risk DCIS a PPV of 0.80, a NPV of 0.84 and an AUC of 0.76 were achieved.

CONCLUSION

For both scenarios (DCIS grade I/II vs. III, DCIS grade I/II vs. III and/or IBC) AUCs were high, 0.72 and 0.76, respectively, concluding that our convolutional neural network can discriminate low-grade from high-grade DCIS.

Prediction of the histologic upgrade of ductal carcinoma in situ using a combined radiomics and machine learning approach based on breast dynamic contrast-enhanced magnetic resonance imaging

Objective

To investigate whether support vector machine (SVM) trained with radiomics features based on breast magnetic resonance imaging (MRI) could predict the upgrade of ductal carcinoma in situ (DCIS) diagnosed by core needle biopsy (CNB) after surgical excision.

Materials and methods

This retrospective study included a total of 349 lesions from 346 female patients (mean age, 54 years) diagnosed with DCIS by CNB between January 2011 and December 2017. Based on histological confirmation after surgery, the patients were divided into pure (n = 198, 56.7%) and upgraded DCIS (n = 151, 43.3%). The entire dataset was randomly split to training (80%) and test sets (20%). Radiomics features were extracted from the intratumor region-of-interest, which was semi-automatically drawn by two radiologists, based on the first subtraction images from dynamic contrast-enhanced T1-weighted MRI. A least absolute shrinkage and selection operator (LASSO) was used for feature selection. A 4-fold cross validation was applied to the training set to determine the combination of features used to train SVM for classification between pure and upgraded DCIS. Sensitivity, specificity, accuracy, and area under the receiver-operating characteristic curve (AUC) were calculated to evaluate the model performance using the hold-out test set.

Results

The model trained with 9 features (Energy, Skewness, Surface Area to Volume ratio, Gray Level Non Uniformity, Kurtosis, Dependence Variance, Maximum 2D diameter Column, Sphericity, and Large Area Emphasis) demonstrated the highest 4-fold mean validation accuracy and AUC of 0.724 (95% CI, 0.619-0.829) and 0.742 (0.623-0.860), respectively. Sensitivity, specificity, accuracy, and AUC using the test set were 0.733 (0.575-0.892) and 0.7 (0.558-0.842), 0.714 (0.608-0.820) and 0.767 (0.651-0.882), respectively.

Conclusion

Our study suggested that the combined radiomics and machine learning approach based on preoperative breast MRI may provide an assisting tool to predict the histologic upgrade of DCIS.

Anomaly Detection of Calcifications in Mammography Based on 11,000 Negative Cases

In mammography, calcifications are one of the most common signs of breast cancer. Detection of such lesions is an active area of research for computer-aided diagnosis and machine learning algorithms. Due to limited numbers of positive cases, many supervised detection models suffer from overfitting and fail to generalize. We present a one-class, semi-supervised framework using a deep convolutional autoencoder trained with over 50,000 images from 11,000 negative-only cases. Since the model learned from only normal breast parenchymal features, calcifications produced large signals when comparing the residuals between input and reconstruction output images. As a key advancement, a structural dissimilarity index was used to suppress non-structural noises. Our selected model achieved pixel-based AUROC of 0.959 and AUPRC of 0.676 during validation, where calcification masks were defined in a semi-automated process. Although not trained directly on any cancers, detection performance of calcification lesions on 1,883 testing images (645 malignant and 1238 negative) achieved 75% sensitivity at 2.5 false positives per image. Performance plateaued early when trained with only a fraction of the cases, and greater model complexity or a larger dataset did not improve performance. This study demonstrates the potential of this anomaly detection approach to detect mammographic calcifications in a semi-supervised manner with efficient use of a small number of labeled images, and may facilitate new clinical applications such as computer-aided triage and quality improvement.

Prediction of Upstaging in Ductal Carcinoma in Situ Based on Mammographic Radiomic Features

Background Improving diagnosis of ductal carcinoma in situ (DCIS) before surgery is important in choosing optimal patient management strategies. However, patients may harbor occult invasive disease not detected until definitive surgery. Purpose To assess the performance and clinical utility of mammographic radiomic features in the prediction of occult invasive cancer among women diagnosed with DCIS on the basis of core biopsy findings. Materials and Methods In this Health Insurance Portability and Accountability Act-compliant retrospective study, digital magnification mammographic images were collected from women who underwent breast core-needle biopsy for calcifications that was performed at a single institution between September 2008 and April 2017 and yielded a diagnosis of DCIS. The database query was directed at asymptomatic women with calcifications without a mass, architectural distortion, asymmetric density, or palpable disease. Logistic regression with regularization was used. Differences across training and internal test set by upstaging rate, age, lesion size, and estrogen and progesterone receptor status were assessed by using the Kruskal-Wallis or χ² test. Results The study consisted of 700 women with DCIS (age range, 40-89 years; mean age, 59 years ± 10 [standard deviation]), including 114 with lesions (16.3%) upstaged to invasive cancer at subsequent surgery. The sample was split randomly into 400 women for the training set and 300 for the testing set (mean ages: training set, 59 years ± 10; test set, 59 years ± 10; P = .85). A total of 109 radiomic and four clinical features were extracted. The best model on the test set by using all radiomic and clinical features helped predict upstaging with an area under the receiver operating characteristic curve of 0.71 (95% CI: 0.62, 0.79). For a fixed high sensitivity (90%), the model yielded a specificity of 22%, a negative predictive value of 92%, and an odds ratio of 2.4 (95% CI: 1.8, 3.2). High specificity (90%) corresponded to a sensitivity of 37%, positive predictive value of 41%, and odds ratio of 5.0 (95% CI: 2.8, 9.0). Conclusion Machine learning models that use radiomic features applied to mammographic calcifications may help predict upstaging of ductal carcinoma in situ, which can refine clinical decision making and treatment planning. © RSNA, 2022.

Ductal Carcinoma in Situ: State-of-the-Art Review

Ductal carcinoma in situ (DCIS) is a nonobligate precursor of invasive cancer, and its detection, diagnosis, and management are controversial. DCIS incidence grew with the expansion of screening mammography programs in the 1980s and 1990s, and DCIS is viewed as a major driver of overdiagnosis and overtreatment. For pathologists, the diagnosis and classification of DCIS is challenging due to undersampling and interobserver variability. Understanding the progression from normal breast tissue to DCIS and, ultimately, to invasive cancer is limited by a paucity of natural history data with multiple proposed evolutionary models of DCIS initiation and progression. Although radiologists are familiar with the classic presentation of DCIS as asymptomatic calcifications at mammography, the expanded pool of modalities, advanced imaging techniques, and image analytics have identified multiple potential biomarkers of histopathologic characteristics and prognosis. Finally, there is growing interest in the nonsurgical management of DCIS, including active surveillance, to reduce overtreatment and provide patients with more personalized management options. However, current biomarkers are not adept at enabling identification of occult invasive disease at biopsy or accurately predicting the risk of progression to invasive disease. Several active surveillance trials are ongoing and are expected to better identify women with low-risk DCIS who may avoid surgery.

Mixed-Methods Study to Predict Upstaging of DCIS to Invasive Disease on Mammography

BACKGROUND. The incidence of ductal carcinoma in situ (DCIS) has steadily increased, as have concerns regarding overtreatment. Active surveillance is a novel treatment strategy that avoids surgical excision, but identifying patients with occult invasive disease who should be excluded from active surveillance is challenging. Radiologists are not typically expected to predict the upstaging of DCIS to invasive disease, though they might be trained to perform this task. OBJECTIVE. The purpose of this study was to determine whether a mixed-methods two-stage observer study can improve radiologists' ability to predict upstaging of DCIS to invasive disease on mammography. METHODS. All cases of DCIS calcifications that underwent stereotactic biopsy between 2010 and 2015 were identified. Two cohorts were randomly generated, each containing 150 cases (120 pure DCIS cases and 30 DCIS cases upstaged to invasive disease at surgery). Nine breast radiologists reviewed the mammograms in the first cohort in a blinded fashion and scored the probability of upstaging to invasive disease. The radiologists then reviewed the cases and results collectively in a focus group to develop consensus criteria that could improve their ability to predict upstaging. The radiologists reviewed the mammograms from the second cohort in a blinded fashion and again scored the probability of upstaging. Statistical analysis compared the performances between rounds 1 and 2. RESULTS. The mean AUC for reader performance in predicting upstaging in round 1 was 0.623 (range, 0.514-0.684). In the focus group, radiologists agreed that upstaging was better predicted when an associated mass, asymmetry, or architectural distortion was present; when densely packed calcifications extended over a larger area; and when the most suspicious features were focused on rather than the most common features. Additionally, radiologists agreed that BI-RADS descriptors do not adequately characterize risk of invasion, and that microinvasive disease and smaller areas of DCIS will have poor prediction estimates. Reader performance significantly improved in round 2 (mean AUC, 0.765; range, 0.617-0.852; p = .045). CONCLUSION. A mixed-methods two-stage observer study identified factors that helped radiologists significantly improve their ability to predict upstaging of DCIS to invasive disease. CLINICAL IMPACT. Breast radiologists can be trained to better predict upstaging of DCIS to invasive disease, which may facilitate discussions with patients and referring providers.

Application of deep learning to predict underestimation in ductal carcinoma in situ of the breast with ultrasound

Background

To develop an ultrasound-based deep learning model to predict postoperative upgrading of pure ductal carcinoma in situ (DCIS) diagnosed by core needle biopsy (CNB) before surgery.

Methods

Of the 360 patients with DCIS diagnosed by CNB and identified retrospectively, 180 had lesions upstaged to ductal carcinoma in situ with microinvasion (DCISM) or invasive ductal carcinoma (IDC) postoperatively. Ultrasound images obtained from the hospital database were divided into a training set (n=240) and validation set (n=120), with a ratio of 2:1 in chronological order. Four deep learning models, based on the ResNet and VggNet structures, were established to classify the ultrasound images into postoperative upgrade and pure DCIS. We obtained the area under the receiver operating characteristic curve (AUROC), specificity, sensitivity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) to estimate the performance of the predictive models. The robustness of the models was evaluated by a 3-fold cross-validation.

Results

Clinical features were not significantly different between the training set and the test set (P value >0.05). The area under the receiver operating characteristic curve of our models ranged from 0.724 to 0.804. The sensitivity, specificity, and accuracy of the optimal model were 0.733, 0.750, and 0.742, respectively. The three-fold cross-validation results showed that the model was very robust.

Conclusions

The ultrasound-based deep learning prediction model is effective in predicting DCIS that will be upgraded postoperatively.

Predicting Upstaging of DCIS to Invasive Disease: Radiologists's Predictive Performance

RATIONALE AND OBJECTIVES

The purpose of this study is to quantify breast radiologists' performance at predicting occult invasive disease when ductal carcinoma in situ (DCIS) presents as calcifications on mammography and to identify imaging and histopathological features that are associated with radiologists' performance.

MATERIALS AND METHODS

Mammographically detected calcifications that were initially diagnosed as DCIS on core biopsy and underwent definitive surgical excision between 2010 and 2015 were identified. Thirty cases of suspicious calcifications upstaged to invasive ductal carcinoma and 120 cases of DCIS confirmed at the time of definitive surgery were randomly selected. Nuclear grade, estrogen and progesterone receptor status, patient age, calcification long axis length, and breast density were collected. Ten breast radiologists who were blinded to all clinical and pathology data independently reviewed all cases and estimated the likelihood that the DCIS would be upstaged to invasive disease at surgical excision. Subgroup analysis was performed based on nuclear grade, long axis length, breast density and after exclusion of microinvasive disease.

RESULTS

Reader performance to predict upstaging ranged from an area under the receiver operating characteristic curve (AUC) of 0.541-0.684 with a mean AUC of 0.620 (95%CI: 0.489-0.751). Performances improved for lesions smaller than 2 cm (AUC: 0.676 vs 0.500; p = 0.002). The exclusion of microinvasive cases also improved performance (AUC: 0.651 vs 0.620; p = 0.005). There was no difference in performance based on breast density (p = 0.850) or nuclear grade (p = 0.270) CONCLUSION: Radiologists were able to predict invasive disease better than chance, particularly for smaller DCIS lesions (<2 cm) and after the exclusion of microinvasive disease.

Prediction of Upstaged Ductal Carcinoma In Situ Using Forced Labeling and Domain Adaptation

OBJECTIVE

The goal of this study is to use adjunctive classes to improve a predictive model whose performance is limited by the common problems of small numbers of primary cases, high feature dimensionality, and poor class separability. Specifically, our clinical task is to use mammographic features to predict whether ductal carcinoma in situ (DCIS) identified at needle core biopsy will be later upstaged or shown to contain invasive breast cancer.

METHODS

To improve the prediction of pure DCIS (negative) versus upstaged DCIS (positive) cases, this study considers the adjunctive roles of two related classes: atypical ductal hyperplasia (ADH), a non-cancer type of breast abnormity, and invasive ductal carcinoma (IDC), with 113 computer vision based mammographic features extracted from each case. To improve the baseline Model A's classification of pure vs. upstaged DCIS, we designed three different strategies (Models B, C, D) with different ways of embedding features or inputs.

RESULTS

Based on ROC analysis, the baseline Model A performed with AUC of 0.614 (95% CI, 0.496-0.733). All three new models performed better than the baseline, with domain adaptation (Model D) performing the best with an AUC of 0.697 (95% CI, 0.595-0.797).

CONCLUSION

We improved the prediction performance of DCIS upstaging by embedding two related pathology classes in different training phases.

SIGNIFICANCE

The three new strategies of embedding related class data all outperformed the baseline model, thus demonstrating not only feature similarities among these different classes, but also the potential for improving classification by using other related classes.

Potential Role of Convolutional Neural Network Based Algorithm in Patient Selection for DCIS Observation Trials Using a Mammogram Dataset

RATIONALE AND OBJECTIVES

We investigated the feasibility of utilizing convolutional neural network (CNN) for predicting patients with pure Ductal Carcinoma In Situ (DCIS) versus DCIS with invasion using mammographic images.

MATERIALS AND METHODS

An IRB-approved retrospective study was performed. 246 unique images from 123 patients were used for our CNN algorithm. In total, 164 images in 82 patients diagnosed with DCIS by stereotactic-guided biopsy of calcifications without any upgrade at the time of surgical excision (pure DCIS group). A total of 82 images in 41 patients with mammographic calcifications yielding occult invasive carcinoma as the final upgraded diagnosis on surgery (occult invasive group). Two standard mammographic magnification views (CC and ML/LM) of the calcifications were used for analysis. Calcifications were segmented using an open source software platform 3D Slicer and resized to fit a 128 × 128 pixel bounding box. A 15 hidden layer topology was used to implement the neural network. The network architecture contained five residual layers and dropout of 0.25 after each convolution. Five-fold cross validation was performed using training set (80%) and validation set (20%). Code was implemented in open source software Keras with TensorFlow on a Linux workstation with NVIDIA GTX 1070 Pascal GPU.

RESULTS

Our CNN algorithm for predicting patients with pure DCIS achieved an overall diagnostic accuracy of 74.6% (95% CI, ±5) with area under the ROC curve of 0.71 (95% CI, ±0.04), specificity of 91.6% (95% CI, ±5%) and sensitivity of 49.4% (95% CI, ±6%).

CONCLUSION

It's feasible to apply CNN to distinguish pure DCIS from DCIS with invasion with high specificity using mammographic images.

Do Eligibility Criteria for Ductal Carcinoma In Situ (DCIS) Active Surveillance Trials Identify Patients at Low Risk for Upgrade to Invasive Carcinoma?

BACKGROUND

Clinical trials are currently ongoing to determine the safety and efficacy of active surveillance (AS) versus usual care (surgical and radiation treatment) for women with ductal carcinoma in situ (DCIS). This study aimed to determine upgrade rates of DCIS at needle biopsy to invasive carcinoma at surgery among women who meet the eligibility criteria for AS trials.

METHODS

A retrospective review was performed of consecutive women at an academic medical center with a diagnosis of DCIS at needle biopsy from 2007 to 2016. Medical records were reviewed for mode of presentation, imaging findings, biopsy pathology results, and surgical outcomes. Each patient with DCIS was evaluated for AS trial eligibility based on published criteria for the COMET, LORD, and LORIS trials.

RESULTS

During a 10-year period, DCIS was diagnosed in 858 women (mean age 58 years; range 28-89 years). Of the 858 women, 498 (58%) were eligible for the COMET trial, 101 (11.8%) for the LORD trial, and 343 (40%) for the LORIS trial. The rates of upgrade to invasive carcinoma were 12% (60/498) for the COMET trial, 5% (5/101) for the LORD trial, and 11.1% (38/343) for the LORIS trial. The invasive carcinomas ranged from 0.2 to 20 mm, and all were node-negative.

CONCLUSIONS

Women who meet the eligibility criteria for DCIS AS trials remain at risk for occult invasive carcinoma at presentation, with upgrade rates ranging from 5 to 12%. These findings suggest that more precise criteria are needed to ensure that women with invasive carcinoma are excluded from AS trials.

Ductal Carcinoma In Situ Biology, Language, and Active Surveillance: A Survey of Breast Radiologists' Knowledge and Opinions

PURPOSE

To understand how breast radiologists perceive ductal carcinoma in situ (DCIS).

MATERIALS AND METHODS

A 19-item survey was developed by the Society of Breast Imaging Patient Care and Delivery Committee and distributed to all Society of Breast Imaging members. The survey queried respondents' demographics, knowledge of DCIS biology, language used to discuss a new diagnosis of DCIS, and perspectives on active surveillance for DCIS. Five-point Likert scales (1 = strongly disagree, 3 = neutral, 5 = strongly agree) were used.

RESULTS

There were 536 responses for a response rate of 41%. There was agreement that DCIS is the primary driver of overdiagnosis in breast cancer screening (median 4), and respondents provided mean and median overdiagnosis estimates of 29.7% and 25% for low-grade DCIS as well as 4.2% and 0% for high-grade DCIS, respectively. Responses varied in how to describe DCIS but most often used the word "cancer" with a qualifier such as "early" (32%) or "pre-invasive" (25%). Respondents disagreed (median 2) with removing the word "carcinoma" from DCIS. Finally, there was agreement that current standard of care therapy for some forms of DCIS is overtreatment (median 4) and that active surveillance as an alternative management strategy should be studied (mean 4), but felt that ultrasound (median 4) and MRI (median 4) should be used to exclude women with occult invasive disease before active surveillance.

CONCLUSIONS

Breast radiologists' opinions about DCIS biology, language, and active surveillance are not homogenous, but general trends exist that can be used to guide research, education, and advocacy efforts.

Ductal Carcinoma in Situ: Current Concepts in Biology, Imaging, and Treatment

Ductal carcinoma in situ (DCIS) of the breast is a group of heterogeneous epithelial proliferations confined to the milk ducts that nearly always present in asymptomatic women on breast cancer screening. A stage 0, preinvasive breast cancer, increased detection of DCIS was initially hailed as a means to prevent invasive breast cancer through surgical treatment with adjuvant radiation and/or endocrine therapies. However, controversy in the medical community has emerged in the past two decades that a fraction of DCIS represents overdiagnosis, leading to unnecessary treatments and resulting morbidity. The imaging hallmarks of DCIS include linearly or segmentally distributed calcifications on mammography or nonmass enhancement on breast MRI. Imaging features have been shown to reflect the biological heterogeneity of DCIS lesions, with recent studies indicating MRI may identify a greater fraction of higher-grade lesions than mammography does. There is strong interest in the surgical, imaging, and oncology communities to better align DCIS management with biology, which has resulted in trials of active surveillance and therapy that is less aggressive. However, risk stratification of DCIS remains imperfect, which has limited the development of precision therapy approaches matched to DCIS aggressiveness. Accordingly, there are opportunities for breast imaging radiologists to assist the oncology community by leveraging advanced imaging techniques to identify appropriate patients for the less aggressive DCIS treatments.

Predicting underestimation of ductal carcinoma in situ: a comparison between radiomics and conventional approaches

PURPOSE

We aimed to investigate the feasibility of predicting invasion carcinoma from ductal carcinoma in situ (DCIS) lesions diagnosed by preoperative core needle biopsy using radiomics signatures, clinical imaging characteristics, and breast imaging reporting and data system (BI-RADS) descriptors on mammography.

METHODS

Retrospectively, we enrolled 362 DCIS patients diagnosed by core needle biopsy, 110 (30.4%) of which had invasive carcinoma confirmed by operation and pathology. We analyzed the images identified suspicious calcification from 250 subjects (161 pure DCIS and 89 DCIS with invasion). A total of 569 calcification radiomics signatures were extracted from microcalcification for each patient. We included a group of routine clinical imaging characteristics and BI-RADS descriptors for comparison purpose. Five feature selection and seven classification methods were evaluated in terms of their prediction performance. We compared the area under the receiver operating characteristic curve (AUC) averaged from tenfold cross-validation of different feature sets to identify the best combination of feature selection and classification methods.

RESULTS

Optimal feature selection and classification methods were identified after evaluating various combinations of feature sets. The best performance was achieved using both radiomics and clinical imaging characteristics (AUC = 0.72) performing better than BI-RADS descriptors or radiomics, but was no significant difference with clinical imaging characteristics (AUC = 0.66). The most significant features found were morphology signatures, first-order statistics, asymmetry/mass prevalence, and nuclear grade.

CONCLUSIONS

We found that the prediction model established using microcalcifications radiomics signatures and clinical imaging characteristics has the potential to identify an understaging of invasive breast cancer.

Prediction of Occult Invasive Disease in Ductal Carcinoma in Situ Using Deep Learning Features

PURPOSE

The aim of this study was to determine whether deep features extracted from digital mammograms using a pretrained deep convolutional neural network are prognostic of occult invasive disease for patients with ductal carcinoma in situ (DCIS) on core needle biopsy.

METHODS

In this retrospective study, digital mammographic magnification views were collected for 99 subjects with DCIS at biopsy, 25 of which were subsequently upstaged to invasive cancer. A deep convolutional neural network model that was pretrained on nonmedical images (eg, animals, plants, instruments) was used as the feature extractor. Through a statistical pooling strategy, deep features were extracted at different levels of convolutional layers from the lesion areas, without sacrificing the original resolution or distorting the underlying topology. A multivariate classifier was then trained to predict which tumors contain occult invasive disease. This was compared with the performance of traditional "handcrafted" computer vision (CV) features previously developed specifically to assess mammographic calcifications. The generalization performance was assessed using Monte Carlo cross-validation and receiver operating characteristic curve analysis.

RESULTS

Deep features were able to distinguish DCIS with occult invasion from pure DCIS, with an area under the receiver operating characteristic curve of 0.70 (95% confidence interval, 0.68-0.73). This performance was comparable with the handcrafted CV features (area under the curve = 0.68; 95% confidence interval, 0.66-0.71) that were designed with prior domain knowledge.

CONCLUSIONS

Despite being pretrained on only nonmedical images, the deep features extracted from digital mammograms demonstrated comparable performance with handcrafted CV features for the challenging task of predicting DCIS upstaging.