Michelle Melisko

BACKGROUND: Automated cardiac image interpretation has the potential to transform clinical practice in multiple ways, including enabling serial assessment of cardiac function by nonexperts in primary care and rural settings. We hypothesized that advances in computer vision could enable building a fully automated, scalable analysis pipeline for echocardiogram interpretation, including (1) view identification, (2) image segmentation, (3) quantification of structure and function, and (4) disease detection. METHODS: Using 14 035 echocardiograms spanning a 10-year period, we trained and evaluated convolutional neural network models for multiple tasks, including automated identification of 23 viewpoints and segmentation of cardiac chambers across 5 common views. The segmentation output was used to quantify chamber volumes and left ventricular mass, determine ejection fraction, and facilitate automated determination of longitudinal strain through speckle tracking. Results were evaluated through comparison to manual segmentation and measurements from 8666 echocardiograms obtained during the routine clinical workflow. Finally, we developed models to detect 3 diseases: hypertrophic cardiomyopathy, cardiac amyloid, and pulmonary arterial hypertension. RESULTS: Convolutional neural networks accurately identified views (eg, 96% for parasternal long axis), including flagging partially obscured cardiac chambers, and enabled the segmentation of individual cardiac chambers. The resulting cardiac structure measurements agreed with study report values (eg, median absolute deviations of 15% to 17% of observed values for left ventricular mass, left ventricular diastolic volume, and left atrial volume). In terms of function, we computed automated ejection fraction and longitudinal strain measurements (within 2 cohorts), which agreed with commercial software-derived values (for ejection fraction, median absolute deviation=9.7% of observed, N=6407 studies; for strain, median absolute deviation=7.5%, n=419, and 9.0%, n=110) and demonstrated applicability to serial monitoring of patients with breast cancer for trastuzumab cardiotoxicity. Overall, we found automated measurements to be comparable or superior to manual measurements across 11 internal consistency metrics (eg, the correlation of left atrial and ventricular volumes). Finally, we trained convolutional neural networks to detect hypertrophic cardiomyopathy, cardiac amyloidosis, and pulmonary arterial hypertension with C statistics of 0.93, 0.87, and 0.85, respectively. CONCLUSIONS: Our pipeline lays the groundwork for using automated interpretation to support serial patient tracking and scalable analysis of millions of echocardiograms archived within healthcare systems.

Importance: Approximately 25% of patients with early-stage breast cancer who receive (neo)adjuvant chemotherapy experience a recurrence within 5 years. Improvements in therapy are greatly needed. Objective: To determine if pembrolizumab plus neoadjuvant chemotherapy (NACT) in early-stage breast cancer is likely to be successful in a 300-patient, confirmatory randomized phase 3 neoadjuvant clinical trial. Design, Setting, and Participants: The I-SPY2 study is an ongoing open-label, multicenter, adaptively randomized phase 2 platform trial for high-risk, stage II/III breast cancer, evaluating multiple investigational arms in parallel. Standard NACT serves as the common control arm; investigational agent(s) are added to this backbone. Patients with ERBB2 (formerly HER2)-negative breast cancer were eligible for randomization to pembrolizumab between November 2015 and November 2016. Interventions: Participants were randomized to receive taxane- and anthracycline-based NACT with or without pembrolizumab, followed by definitive surgery. Main Outcomes and Measures: The primary end point was pathologic complete response (pCR). Secondary end points were residual cancer burden (RCB) and 3-year event-free and distant recurrence-free survival. Investigational arms graduated when demonstrating an 85% predictive probability of success in a hypothetical confirmatory phase 3 trial. Results: Of the 250 women included in the final analysis, 181 were randomized to the standard NACT control group (median [range] age, 47 [24.77] years). Sixty-nine women (median [range] age, 50 [27-71] years) were randomized to 4 cycles of pembrolizumab in combination with weekly paclitaxel followed by AC; 40 hormone receptor (HR)-positive and 29 triple-negative. Pembrolizumab graduated in all 3 biomarker signatures studied. Final estimated pCR rates, evaluated in March 2017, were 44% vs 17%, 30% vs 13%, and 60% vs 22% for pembrolizumab vs control in the ERBB2-negative, HR-positive/ERBB2-negative, and triple-negative cohorts, respectively. Pembrolizumab shifted the RCB distribution to a lower disease burden for each cohort evaluated. Adverse events included immune-related endocrinopathies, notably thyroid abnormalities (13.0%) and adrenal insufficiency (8.7%). Achieving a pCR appeared predictive of long-term outcome, where patients with pCR following pembrolizumab plus chemotherapy had high event-free survival rates (93% at 3 years with 2.8 years' median follow-up). Conclusions and Relevance: When added to standard neoadjuvant chemotherapy, pembrolizumab more than doubled the estimated pCR rates for both HR-positive/ERBB2-negative and triple-negative breast cancer, indicating that checkpoint blockade in women with early-stage, high-risk, ERBB2-negative breast cancer is highly likely to succeed in a phase 3 trial. Pembrolizumab was the first of 10 agents to graduate in the HR-positive/ERBB2-negative signature. Trial Registration: ClinicalTrials.gov Identifier: NCT01042379.

BACKGROUND: The post-treatment quality of life (QOL) impacts of receiving precancer-treatment infertility counseling and of pursuing fertility preservation have not been described in large-scale studies of reproductive age women with cancer. METHODS: In total, 1041 women who were diagnosed between ages 18 and 40 years responded to a retrospective survey and reported whether they received infertility counseling before cancer treatment and whether they took action to preserve fertility. Five cancer types were included: leukemia, Hodgkin disease, non-Hodgkin lymphoma, breast cancer, and gastrointestinal cancer. Validated QOL scales were used: the Decision Regret Score, the Satisfaction with Life Scale (SWLS), and the brief World Health Organization QOL questionnaire. RESULTS: Overall, 560 women (61%) who received treatment that potentially could affect fertility were counseled by the oncology team, 45 (5%) were counseled by fertility specialists, and 36 (4%) took action to preserve fertility. Pretreatment infertility counseling by a fertility specialist and an oncologist resulted in lower regret than counseling by an oncologist alone (8.4 vs 11.0; P < .0001). The addition of fertility preservation (6.6 vs 11.0; P < .0001) also was associated with even lower regret scores than counseling by an oncologist alone. Further improvements also were observed in SWLS scores with the addition of fertility specialist counseling (23.0 vs 19.8; P = .09) or preserving fertility (24.0 vs 19.0; P = .05). CONCLUSIONS: Receiving specialized counseling about reproductive loss and pursuing fertility preservation is associated with less regret and greater QOL for survivors, yet few patients are exposed to this potential benefit. Women of reproductive age should have expert counseling and should be given the opportunity to make active decisions about preserving fertility.

PURPOSE Olaparib, a poly (ADP-ribose) polymerase (PARP) inhibitor (PARPi), is approved for the treatment of human epidermal growth factor receptor 2 (HER2)–negative metastatic breast cancer (MBC) in germline (g) BRCA1/ 2 mutation carriers. Olaparib Expanded, an investigator-initiated, phase II study, assessed olaparib response in patients with MBC with somatic (s) BRCA1/ 2 mutations or g/s mutations in homologous recombination (HR)–related genes other than BRCA1/2. METHODS Eligible patients had MBC with measurable disease and germline mutations in non- BRCA1/ 2 HR-related genes (cohort 1) or somatic mutations in these genes or BRCA1/ 2 (cohort 2). Prior PARPi, platinum-refractory disease, or progression on more than two chemotherapy regimens (metastatic setting) was not allowed. Patients received olaparib 300 mg orally twice a day until progression. A single-arm, two-stage design was used. The primary endpoint was objective response rate (ORR); the null hypothesis (≤ 5% ORR) would be rejected within each cohort if there were four or more responses in 27 patients. Secondary endpoints included clinical benefit rate and progression-free survival (PFS). RESULTS Fifty-four patients enrolled. Seventy-six percent had estrogen receptor–positive HER2-negative disease. Eighty-seven percent had mutations in PALB2, s BRCA1/ 2, ATM, or CHEK2. In cohort 1, ORR was 33% (90% CI, 19% to 51%) and in cohort 2, 31% (90% CI, 15% to 49%). Confirmed responses were seen only with g PALB2 (ORR, 82%) and s BRCA1/ 2 (ORR, 50%) mutations. Median PFS was 13.3 months (90% CI, 12 months to not available/computable [NA]) for g PALB2 and 6.3 months (90% CI, 4.4 months to NA) for s BRCA1/ 2 mutation carriers. No responses were observed with ATM or CHEK2 mutations alone. CONCLUSION PARP inhibition is an effective treatment for patients with MBC and g PALB2 or s BRCA1/ 2 mutations, significantly expanding the population of patients with breast cancer likely to benefit from PARPi beyond g BRCA1/ 2 mutation carriers. These results emphasize the value of molecular characterization for treatment decisions in MBC.

BACKGROUND: The heterogeneity of breast cancer makes identifying effective therapies challenging. The I-SPY 2 trial, a multicenter, adaptive phase 2 trial of neoadjuvant therapy for high-risk clinical stage II or III breast cancer, evaluated multiple new agents added to standard chemotherapy to assess the effects on rates of pathological complete response (i.e., absence of residual cancer in the breast or lymph nodes at the time of surgery). METHODS: We used adaptive randomization to compare standard neoadjuvant chemotherapy plus the tyrosine kinase inhibitor neratinib with control. Eligible women were categorized according to eight biomarker subtypes on the basis of human epidermal growth factor receptor 2 (HER2) status, hormone-receptor status, and risk according to a 70-gene profile. Neratinib was evaluated against control with regard to 10 biomarker signatures (prospectively defined combinations of subtypes). The primary end point was pathological complete response. Volume changes on serial magnetic resonance imaging were used to assess the likelihood of such a response in each patient. Adaptive assignment to experimental groups within each disease subtype was based on Bayesian probabilities of the superiority of the treatment over control. Enrollment in the experimental group was stopped when the 85% Bayesian predictive probability of success in a confirmatory phase 3 trial of neoadjuvant therapy reached a prespecified threshold for any biomarker signature ("graduation"). Enrollment was stopped for futility if the probability fell to below 10% for every biomarker signature. RESULTS: Neratinib reached the prespecified efficacy threshold with regard to the HER2-positive, hormone-receptor-negative signature. Among patients with HER2-positive, hormone-receptor-negative cancer, the mean estimated rate of pathological complete response was 56% (95% Bayesian probability interval [PI], 37 to 73%) among 115 patients in the neratinib group, as compared with 33% among 78 controls (95% PI, 11 to 54%). The final predictive probability of success in phase 3 testing was 79%. CONCLUSIONS: Neratinib added to standard therapy was highly likely to result in higher rates of pathological complete response than standard chemotherapy with trastuzumab among patients with HER2-positive, hormone-receptor-negative breast cancer. (Funded by QuantumLeap Healthcare Collaborative and others; I-SPY 2 TRIAL ClinicalTrials.gov number, NCT01042379.).