Citations

Purpose This study aimed to compare accuracy and efficiency of a convolutional neural network (CNN)–enhanced workflow for pancreas segmentation versus radiologists in the context of interreader reliability. Methods Volumetric pancreas segmentations on a data set of 294 portal venous computed tomographies were performed by 3 radiologists (R1, R2, and R3) and by a CNN. Convolutional neural network segmentations were reviewed and, if needed, corrected (“corrected CNN [c-CNN]” segmentations) by radiologists. Ground truth was obtained from radiologists' manual segmentations using simultaneous truth and performance level estimation algorithm. Interreader reliability and model's accuracy were evaluated with Dice-Sorenson coefficient (DSC) and Jaccard coefficient (JC). Equivalence was determined using a two 1-sided test. Convolutional neural network segmentations below the 25th percentile DSC were reviewed to evaluate segmentation errors. Time for manual segmentation and c-CNN was compared. Results Pancreas volumes from 3 sets of segmentations (manual, CNN, and c-CNN) were noninferior to simultaneous truth and performance level estimation–derived volumes [76.6 cm3 (20.2 cm3), P < 0.05]. Interreader reliability was high (mean [SD] DSC between R2-R1, 0.87 [0.04]; R3-R1, 0.90 [0.05]; R2-R3, 0.87 [0.04]). Convolutional neural network segmentations were highly accurate (DSC, 0.88 [0.05]; JC, 0.79 [0.07]) and required minimal-to-no corrections (c-CNN: DSC, 0.89 [0.04]; JC, 0.81 [0.06]; equivalence, P < 0.05). Undersegmentation (n = 47 [64%]) was common in the 73 CNN segmentations below 25th percentile DSC, but there were no major errors. Total inference time (minutes) for CNN was 1.2 (0.3). Average time (minutes) taken by radiologists for c-CNN (0.6 [0.97]) was substantially lower compared with manual segmentation (3.37 [1.47]; savings of 77.9%–87% [P < 0.0001]). Conclusions Convolutional neural network–enhanced workflow provides high accuracy and efficiency for volumetric pancreas segmentation on computed tomography.

AbstractBackground:Triple-negative breast cancer (TNBC) is the most aggressive breast cancer subtype. Patients with TNBC are primarily treated with neoadjuvant chemotherapy (NAC). Response to NAC is prognostic, with reductions in low overall survival and disease-free survival rates in those patients who do not achieve pathological complete response (pCR). Based on this premise, we focused on patients with residual disease and hypothesized that paired analysis of primary and residual TNBC tumors following NAC could identify unique biomarkers associated with post-NAC recurrence.Methods and Results: We investigated 24 TNBC tumors from 12 patients with pre- and post-NAC data, including four patients who recurred shortly after surgery(<24 months) and eight who remained recurrence-free (> 48 months). These tumors were collected from a prospective NAC breast cancer study (BEAUTY) conducted at the Mayo Clinic. Differential expression analysis of pre-NAC biopsies showed minimal gene expression differences between early recurrent and non-recurrent TNBC tumors; however, post-NAC samples demonstrated a significant alteration in expression patterns in response to intervention. Topological-level differences associated with early recurrence were implicated in 251 gene sets, and an independent assessment of microarray gene expression data from the paired NAC I-SPY1 trial confirmed 56 genesets. Underlying these 56 gene sets, 113 genes were observed to be differentially expressed in the I-SPY1 and BEAUTY post-NAC studies. An independent (n=392) breast cancer data set with relapse-free survival (RFS) data was used to refine our gene list to a 17 gene signature. A three-fold cross-validation analysis of the gene signature with the combined BEAUTY and I-SPY1 data yielded an average AUC of 0.88 for six machine learning models. Due to the limited studies with pre-and post-NAC TNBC tumor data, further validation of the signature is needed.Conclusion:Analysis of multi-omics data from post-NAC TNBC chemoresistant tumors showed downregulation of mismatch repair, purin, and tubulin pathways. Additionally, weidentified a 17 gene signature in TNBC associated with post-NAC recurrence enriched with down-regulation of immune genes.

Abstract Purpose: To evaluate prediction of response and event-free survival (EFS) following neoadjuvant endocrine therapy by SET2,3 index of nonproliferation gene expression related to estrogen and progesterone receptors adjusted for baseline prognosis. Experimental Design: A correlative study was conducted of SET2,3 measured from gene expression profiles of diagnostic tumor (Agilent microarrays) in 379 women with cStage II–III breast cancer from the American College of Surgeons Oncology Group Z1031 neoadjuvant aromatase inhibitor trial SET2,3 was dichotomized using the previously published cutoff. Fisher exact test was used to assess the association between SET2,3 and low proliferation at week 2–4 [Ki67 ≤ 10% or complete cell-cycle arrest (CCCA; Ki67 ≤ 2.7%)] and PEPI-0 rate in cohort B, and the association between SET2,3 and ypStage 0/I in all patients. Cox models were used to assess EFS with respect to SET2,3 excluding cohort B patients who switched to chemotherapy. Results: Patients with high SET2,3 had higher rate of pharmacodynamic response than patients with low SET2,3 (Ki67 ≤ 10% in 88.2% vs. 56.9%, P &lt; 0.0001; CCCA in 50.0% vs. 26.2%, P = 0.0054), but rate of ypStage 0/I (24.0% vs. 20.4%, P = 0.4580) or PEPI = 0 (28.4% vs. 20.6%, P = 0.3419) was not different. Patients with high SET2,3 had longer EFS than patients with low SET2,3 (HR, 0.52, 95% confidence interval: 0.34–0.80; P = 0.0026). Conclusions: This exploratory analysis of Z1031 data demonstrated a higher rate of pharmacodynamic suppression of proliferation and longer EFS in high SET2,3 disease relative to low SET2,3 disease. The ypStage 0/I rate and PEPI = 0 rate were similar with respect to SET2,3.

Abstract Triple-negative breast cancer (TNBC) is the most aggressive breast cancer subtype with low overall survival rates and high molecular heterogeneity; therefore, few targeted therapies are available. The luminal androgen receptor (LAR) is the most consistently identified TNBC subtype, but the clinical utility has yet to be established. Here, we constructed a novel genomic classifier, LAR-Sig, that distinguishes the LAR subtype from other TNBC subtypes and provide evidence that it is a clinically distinct disease. A meta-analysis of seven TNBC datasets (n = 1086 samples) from neoadjuvant clinical trials demonstrated that LAR patients have significantly reduced response (pCR) rates than non-LAR TNBC patients (odds ratio = 2.11, 95% CI: 1.33, 2.89). Moreover, deconvolution of the tumor microenvironment confirmed an enrichment of luminal epithelium corresponding with a decrease in basal and myoepithelium in LAR TNBC tumors. Increased immunosuppression in LAR patients may lead to a decreased presence of cycling T-cells and plasma cells. While, an increased presence of myofibroblast-like cancer-associated cells may impede drug delivery and treatment. In summary, the lower levels of tumor infiltrating lymphocytes (TILs), reduced immune activity in the micro-environment, and lower pCR rates after NAC, suggest that new therapeutic strategies for the LAR TNBC subtype need to be developed.

Abstract Introduction: The ALTERNATE trial randomized postmenopausal women with ER Allred 6-8 HER2- breast cancer to 6 months of NET with anastrozole (A), fulvestrant (F) or the combination (A+F). Biopsies were taken preNET and after 4-weeks(wks). Patients with Ki67 values &gt;10% at 4-wks were offered triage to neoadjuvant chemotherapy. Patients with on-treatment Ki67 ≤ 10% who completed NET underwent surgery and Ki67 was reassessed. The primary endpoint was endocrine-sensitive disease rate (ESDR). ESD is defined as pCR or PEPI-0 residual disease (pT1-2, pN0, Ki67 ≤ 2.7%). We previously reported that the ESDR difference between the F-containing arms and the A arm was not &gt;10% (ASCO 2020) and that baseline RNA-seq-based intrinsic subtypes predicted outcomes overall (SABCS 2021). Herein we describe relationships between PAM50 intrinsic subtype and Ki67 values by treatment arm because comparative drug effectiveness in adjuvant endocrine therapy studies in ER+ HER2- breast cancer can be predicted by the degree of Ki67 suppression (PMC3518447). Methods: 743 of the 1297 eligible patients (A: 264; F: 231; A+F: 248) had RNA extracted from preNET frozen tumor biopsies with &gt;50% tumor content and subjected to RNA seq. Intrinsic subtypes were then assigned as LumA, LumB, and NonLum (Basal or HER2-E) using open-source PAM50-based informatics. Differences in the proportion with wk4 Ki67 &gt; 10%, % change in wk4 ki67, and surgical CCCA (Ki67 ≤ 2.7%) rate (sxCCCA) between treatments and by intrinsic subtype was assessed using stratified logistic regression, Wilcoxon rank sum test, and Fisher’s exact test, respectively. Analysis of sxCCCA excluded those who failed to complete NET for reasons other than disease progression or early Ki67 &gt;10%. Results: Amongst the 358 LumA cases there were no significant differences in Ki67-based endpoints between treatments. Among the 292 LumB cases, the wk4 ki67 &gt; 10% rate was lower with A+F (19.4%) than A (43%) (P=0.0002) and was somewhat lower in F (31%) versus A (P=0.076). The % change in wk4 Ki67 in LumB cases, adjusted for baseline Ki67, showed markedly superior suppression for A+F versus A (-90% vs. -77%; P=&lt;0.0001) and versus F (-90% vs. -80%; P=0.0026). Furthermore sxCCCA rates were significantly higher with A+F than A (53% vs. 25% P = &lt;0.0001) and somewhat higher for F (37%) than A (p=0.068), indicating that superior antiproliferative effects for A+F persist after 6 months on therapy. Lack of Ki67 suppression in response to treatment was observed in the majority of 43 NonLum samples regardless of treatment. Conclusion: The combination of A+F was significantly more effective than either drug alone for the control of LumB breast cancer cell proliferation. This suggests that A+F may be a more effective adjuvant endocrine therapy than A alone in LumB disease. The lower Ki67 suppression with A alone also suggests that poorer outcome in some LumB tumors may be due to insufficient ER targeting rather than ER-independent tumor growth Support: U10CA180821, U10CA180882, U24CA196171, UG1CA189856, U10CA180868 (NRG), NCI BIQSFP, BCRF, Genentech, AstraZeneca. https://acknowledgments.alliancefound.org. (MJE) CPRIT RR140033, P50-CA186784, P50-CA58223, U01-CA214125, U24-CA210954, Gift from Ralph and Lisa Eads, McNair Scholarship. ClinicalTrials.gov Identifier: NCT01953588 Citation Format: Matthew J. Ellis, Meenakshi Anurag, Jeremy Hoog, Aranzazu Fernandez-Martinez, Cheng Fan, Richard Gibbs, Souzan Sanati, Kiran Vij, Mark Watson, Travis Dockter, Olwen Hahn, Joseph Guenther, Abigail Caudle, Erica Crouch, Amy Tiersten, Monica Mita, Wajeeha Razaq, Tina J. Hieken, Yang Wang, A. Marilyn Leitch, Gary W. Unzeitig, Eric Winer, Anna Weiss, Kelly Hunt, Ann H. Partridge, Charles M. Perou, Vera Suman, Cynthia X. Ma, Lisa A. Carey. The effect of intrinsic subtype on inhibition of tumor growth by anastrozole vs. fulvestrant vs. the combination: Results from the Alliance neoadjuvant endocrine therapy (NET) ALTERNATE trial [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr CT026.

Abstract Background: Neoadjuvant ET (NET) offers an opportunity to assess ET sensitivity for ER+ HER2- BC and potentially to tailor therapy. Ki67 &gt;10% on biopsy after 2-4 weeks (wks) of NET identifies patients (pts) with intrinsic ET resistance; while pathologic complete response (pCR) and modified preoperative endocrine prognostic index of 0 (mPEPI 0: pT1-2N0, Ki67 ≤2.7%) at surgery indicates sensitivity to ET. However, on-NET biopsy is not always acceptable or feasible and delays the ET sensitivity determination. PAM50 ROR score and intrinsic subtypes by tumor RNA profiling are prognostic in pts with early stage ER+ HER2- BC, and predict pCR rates to neoadjuvant chemotherapy (NCT) (PMC2667820). We therefore hypothesized that PAM50 analysis on pre-NET biopsies could predict the likelihood of a) a high on-NET Ki67, b) mPEPI-0 or pCR at surgery and, c) pCR for pts triaged to NCT. Methods: The ALTERNATE trial is a phase III study that randomized postmenopausal pts with clinical stage II/III ER+ (Allred score 6-8) HER2- BC to receive neoadjuvant anastrozole, fulvestrant, or both for 6 months before surgery. Research biopsy was required at pre-NET and wk 4, then optional at wk 12. Pts with Ki67 &gt;10% on biopsy at wk 4 or 12 discontinued NET and were offered NCT. PAM50 intrinsic subtype and ROR-P values were generated from mRNA sequencing (RNASeq) analysis on pre-NET biopsies using open-source informatics (PMC7723687) and evaluated for prediction of on-NET Ki67 &gt;10% at wk 4 or 12, pCR or mPEPI-0 post NET, and pCR post NCT. Results: 749 of 1,297 eligible trial pts were included in the analyses, after excluding 548 pts due to insufficient pre-NET tumor for RNASeq (n=511) or PAM50 normal subtype (n=37). Similar to the entire ALTERNATE population, the rate of Ki67 &gt;10% at wk 4 or 12 was 24.4% (95% CI: 21.4-27.7%) and the rate of mPEPI-0/pCR post NET was 19.8% (95% CI: 17.0-22.8%). There were 393 (52.5%) Lum A, 302 (40.3%) Lum B, and 54 (7.2%) non-Lum (9 Basal, 45 HER2-E) BCs. These included 196 (26.2%) ROR-P low, 354 (47.3%) ROR-P medium and 199 (26.6%) ROR-P high BCs. Both the rates of Ki67 &gt;10% at wk 4 or 12 and mPEPI-0/pCR differed significantly with respect to PAM50 subtype or ROR-P category, such that Lum A or ROR-P low BCs were least likely to have a Ki67 &gt;10% at wk 4 or 12 and most likely to achieve mPEPI-0/pCR (Table). 93 of 168 (55.4%) pts triaged to NCT had RNA-seq results, yielding 26 Lum A, 49 Lum B, 4 Basal and 14 HER2-E, with the pCR rates of 0%, 6.1%, 0%, and 21.4%, respectively. There were 10 ROR-P low, 39 medium, and 44 high tumors, with a pCR rate of 0%, 5.1% and 9.1%, respectively. Conclusion: These data indicate that both baseline ROR-P and intrinsic subtype are predictive of early on-NET Ki67 &gt; 10% and mPEPI 0/pCR at surgery after NET. For pts triaged to NCT based on an early on-NET Ki67 &gt;10%, the HER2-E group had the highest pCR rate (20%) and no pCRs were observed in Lum A. These data may be useful for directing neoadjuvant therapy in postmenopausal pts with ER+ HER2- BC. Support: U10CA180821, U10CA180882, U24CA196171, UG1CA189856, U10CA180868 (NRG), NCI BIQSFP, BCRF, Genentech, AstraZeneca. https://acknowledgments.alliancefound.org. (MJE) CPRIT RR140033, P50CA186784, P50-CA58223, U01 CA214125, U24CA210954, Gift from Ralph and Lisa Eads, McNair Scholarship. Trials.gov Identifier: NCT01953588. Table 1.Rates of Ki67 &gt;10% and mPEPI-0/pCR post NET by PAM50 subtype and ROR-P categoryKi67 &gt;10% at wk 4 or 12mPEPI 0/pCR post NETPAM50 SubtypenYes, n (%)PnNo, n (%)PLum A37251 (13.7%) 95% CI: 10.4-17.6%&lt;0.0001393104 (26.5%) 95%CI: 22.2-31.1%&lt;0.0001Lum B29394 (32.1%) 95% CI: 26.8-37.8%30243 (14.2%) 95%CI: 10.5-18.7%Non-luminal (Basal and HER2-E)5338 (71.7%) 95%CI: 57.6-83.2%541 (1.9%) 95%CI: 0.05-9.9%ROR-P CategorynYes, n (%)PnNo, n (%)PLow18018 (10.0%) 95%CI: 6.0-15.3%&lt;0.000119660 (30.6%) 95%CI: 24.2-37.6%&lt;0.0001Intermediate34474 (21.5%) 95%CI: 17.3-26.2%35471 (20.1%) 95%CI: 16.0-24.6%High19491 (46.9%) 95%CI: 39.7-54.2%19917 (8.5%) 95%CI: 5.1-13.3% Citation Format: Cynthia X Ma, Meenakshi Anurag, Travis Dockter, Jeremy Hoog, Aranzazu Fernandez-Martinez, Cheng Fan, Richard Gibbs, Souzan Sanati, Kiran Vij, Mark Watson, Olwen Hahn, Joseph Guenther, Abigail Caudle, Erika Crouch, Amy Tiersten, Monica Mita, Wajeeha Razaq, Tina J Hieken, Yang Wang, A. Marilyn Leitch, Gary W Unzeitig, Anna Weiss, Eric P Winer, Kelly Hunt, Ann H Partridge, Lisa A Carey, Charles M Perou, Matthew J Ellis, Vera Suman. Pam50 intrinsic subtype and risk of recurrence score (ROR) for the prediction of endocrine (ET) sensitivity and pathologic response to chemotherapy in postmenopausal women with clinical stage II/III estrogen receptor positive (ER+) and HER2 negative (HER2-) breast cancer (BC) in the alternate trial (Alliance A011106) [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD9-03.