Abstract 464: AI-powered segmentation and analysis of nuclei morphology predicts genomic and clinical markers in multiple cancer types

5536 Background: Immune checkpoint blockade (ICB) is being explored as a treatment option in ovarian cancer, but objective response rates for single agent ICB are modest at around 10-15%. Validated biomarkers are needed to predict which patients will respond to ICB. BRCA mutations and homologous recombination deficiency (HRD) status are the only validated integral biomarkers in ovarian cancer. HRD tumors exhibit defective DNA repair mechanisms that promote increased mutational burden, which we postulate may correlate with higher neoantigen load and increased expression of targetable immune checkpoints. Methods: The Cancer Genome Atlas (TCGA) ovarian cancer dataset was evaluated and previously published, well annotated samples were obtained for HRD status. HLA type was determined with OptiType. Nonsynonymous mutations were annotated with Ensembl VEP. pVAC-Seq using NetMHCpan algorithm predicted neoepitopes 9 amino acids in length for MHC class I, reporting only those with a predicted IC50 less than 500 nM. Immune checkpoint gene expression counts were normalized with TCGAbiolinks. Correlation between HRD status and neoantigen load was assessed by Wilcoxon test. After log2 transformation, Wilcoxon tests evaluated for association between HRD status and expression of immune checkpoints. The relationship between HRD status and PD-L1 protein abundance with reverse phase protein array was measured. Results: Data from 154 HRD positive and 198 HRD negative tumors were analyzed. HRD positive status correlated with higher neoantigen load (p = 0.038) and increased expression of the immune checkpoints CTLA4 (p = 0.024), TIGIT (p = 0.027), and PVR (p = 0.002), but not PD-L1 (p = 0.238), LAG3 (p = 0.583), HVEM (p = 0.805), GAL9 (p = 0.750), NECTIN2 (p = 0.874), VSIG3 (p = 0.438), PSGL1 (p = 0.205) or VISTA (p = 0.531). TIM3 (p = 0.064) and B7H3 (p = 0.052) both demonstrated a trend towards increased expression in HRD tumors. Interestingly, HRD status showed a negative association with PVRIG (p = 0.028). There was no association between PD-L1 protein abundance and HRD status. Conclusions: HRD positive ovarian tumors demonstrate higher neoantigen load than HRD negative tumors, as well as increased expression of certain immune checkpoints. This supports the hypothesis that increased neoantigen load leads to compensatory induction of immune checkpoints, and suggests that HRD status may predict response to ICB, particularly to drugs that target CTLA4, TIGIT, PVR, TIM3 and B7H4.

Introduction/Aims: DDIR signature, HRD, and stromal tumor infiltrating lymphocytes (sTIL) have each been associated with favorable outcomes in early stage TNBC. We assessed the overlap between these markers and created prognostic categories based on their combined use in a prospective trial of TNBC patients uniformly treated with adjuvant doxorubicin (A) and cyclophosphamide (C) on the SWOG S9313 trial. Methods: SWOG S9313 trial accrued 3,125 women with early stage breast cancer to two alternative dose schedules of AC, with no difference in outcomes between the two arms. 425 centrally determined TNBC cases from S9313 were identified. DDIR signature (score ≥ 0.3681 = DDIR+, Almac Diagnostic Services), HRD status (score ≥ 42 = HRD+, Myriad Genetics), and sTIL were assessed. Gene expression data (Xcel™ array) was subjected to claraT V3.0.0 biological signature analysis (Almac Diagnostic Services), and co-expression cluster analysis was used to identify signatures associated with DDIR and HRD status. The impact of dual classification by DDIR and HRD status (Group 1: DDIR+/HRD+, Group 2: DDIR+/HRD-, Group 3: DDIR-/HRD+, Group 4: DDIR-/HRD-) on disease free survival (DFS) and overall survival (OS) was examined using Cox regression with adjustment for randomized treatment assignment and nodal status. Results: For the 425 patients, median age was 45 years, 33% were node-positive, and 5-year DFS and OS were 74% and 82%, respectively. DDIR and HRD status was available for 89% each, sTIL% was available for 99%, and all three markers were available for 77% (328/425) of patients. 60% were DDIR+ and 65% HRD+. Among DDIR- tumors, 58% were HRD+. sTIL% was associated with DDIR status (P<0.0001) but not with HRD status (P=0.75). The proportion of patients in each group, median sTIL%, and 5-year DFS and OS for each group are outlined in Table 1. DFS and OS were similar for Groups 1, 2, and 3 but significantly lower for Group 4. As expected, cluster analysis showed that immune response signatures dominated Groups 1 and 2 regardless of HRD status. Group 3 tumors were characterized by over-representation of genomic instability signatures, a paucity of immune-related signatures, and low sTIL infiltration. Despite this immune-depleted phenotype, the 5-year OS for Group 3 was similar to that of the immune-enriched DDIR+ groups. Signatures associated with epithelial-mesenchymal transition, mast cell infiltration, and xenobiotic metabolism were over-represented in Group 4. Conclusions: Forty percent of patients with early stage TNBC demonstrate immune-deplete (DDIR-) phenotype, and within this phenotype, more than half demonstrate HRD+ status. HRD+ status within the immune-deplete phenotype predicts for better DFS and OS with adjuvant AC, probably due to underlying genomic instability and increased sensitivity to DNA damaging chemotherapy. Sixty percent of early stage TNBC patients demonstrate an immune-enriched (DDIR+) phenotype, and this phenotype is associated with improved survival with adjuvant AC chemotherapy regardless of HRD status. These findings provide important insights for patient selection and stratification in ongoing and future trials assessing DNA damaging therapy (e.g. PARPi, anthracyclines, platinum agents), immunotherapy, and their combinations in TNBC. Table 1Immune-Enriched GroupsImmune-Deplete GroupsGroup 1 DDIR+/HRD+ N=137 (42%)Group 2 DDIR+/HRD- N=59 (18%)Group 3 DDIR-/HRD+ N=77 (23%)Group 4 DDIR-/HRD- N=55 (17%)5-year DFS82%74%74%56%P=0.001 (Group 1 vs 4); P=0.006 (Group 2 vs 4); P=0.016 (Group 3 vs 4); P=NS (Groups 1 vs 2, 1 vs 3, and 2 vs 3)5-year OS88%86%83%69%P=0.001 (Group 1 vs 4); P=0.003 (Group 2 vs 4); P=0.026 (Group 3 vs 4); P=NS (Group 1 vs 2, 1 vs 3, and 2 vs 3)Median sTIL (%)20%20%5%5%P<0.0001 (Group 1 vs 3); P<0.0001 (Group 2 vs 3); P=NS (Group 1 vs 2, 3 vs 4) Citation Format: Shane R Stecklein, William Barlow, Lajos Pusztai, Kirsten Timms, Richard Kennedy, Sunil Badve, Yesim Gökmen-Polar, Peggy Porter, Hannah Linden, Debu Tripathy, Gabriel N Hortobagyi, Andrew K Godwin, Alastair Thompson, Daniel Hayes, Priyanka Sharma. Classification of triple negative breast cancer (TNBC) by DNA damage immune response (DDIR) signature and homologous recombination deficiency (HRD) status: Analysis of SWOG S9313 adjuvant trial [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr GS3-05.

Background: The combination of cediranib (ced) and olaparib (olap) improves progression-free survival and overall response rates (ORR) in women with recurrent plat sensitive high-grade serous (HGS) ovarian cancer compared to olap alone. In this Phase 2 study of ced/olap in women with relapsed ovarian cancer (NCT02345265), we observed an ORR of 74% with ced/olap in plat sensitive relapsed patients (pts) and an ORR of 20% in plat resistant pts. BROCA-HR was performed to characterize tumor HRD status and to correlate this with the activity of ced/olap in the plat sensitive and resistant settings. Methods: Pts with relapsed ovarian cancer across 11 centers were enrolled to 2 cohorts, either plat sensitive or resistant. All pts received ced 30mg daily and olap tablets 200mg BID. Archival, pre-treatment, and on-treatment biopsies were collected on all pts where safely feasible; pts also had the option of undergoing a post-progression biopsy. A baseline blood sample was collected for germline testing and matched control. Available samples were sequenced with BROCA-HR, and HRD status was determined by (1) presence of an HR-related mutation or (2) an increase in loss of heterozygosity (LOH) score. Results: 72 pts were enrolled, with 70 pts receiving study treatment (35 plat sensitive and 35 plat resistant). BROCA-HR analysis is complete on 60 samples, with 10 samples in final analysis. The concordance of HRD status by presence of an HR-related mutation or by LOH score was relatively low (agreement=36/60=60%, Cohen's Kappa=0.28). In 19 pts with a germline BRCA mutation, 16 were HRD+ by LOH (LOH high), 2 were HRD- by LOH (LOH low), and 1 not quantifiable. 3 pts had a somatic BRCA mutation; 1 was LOH high, 1 LOH low, and 1 not quantifiable. 4 pts had non-BRCA somatic HR-related mutations (3 CDK12; 1 ATR); 3 were LOH high and 1 LOH low. Of the 34 pts without HR-related mutations, 15 were LOH high, 16 were LOH low, and 3 not quantifiable. In 29 plat sensitive pts with BROCA-HR results, 18 were HRD+ by mutation (12 germline BRCA, 3 somatic BRCA, 3 somatic other), and 16 were HRD+ by LOH (LOH high). The ORR in plat sensitive pts by HRD status with available HRD results was: 22/29=75.9% (13/18=72% HRDmut+; 9/11=82% HRDmut-); (12/16=75% LOH high; 6/9= 67% LOH low). In 31 plat resistant pts with BROCA-HR results, 8 were HRD+ by mutation (7 germline BRCA, 0 somatic BRCA, 1 somatic other), and 19 were HRD+ by LOH. The ORR in plat resistant pts by HRD status with available HRD results was: 7/31=22.6% (4/8=50% HRDmut+; 3/23=13% HRDmut-, P=0.05); (7/19=37% LOH high; 0/11= 0% LOH low, P=0.03). Updated results with the full data set are pending. Conclusion: Ced/olap has clinical activity in women with relapsed ovarian cancer, both in the plat sensitive and plat resistant settings. In women with plat sensitive disease, where response rates are high, HRD status by presence of an HR-related mutation or LOH status does not appear to provide significant additional information regarding likelihood of response. In women with plat resistant disease, the presence of HRD by an HR-related mutation or LOH status is correlated with a higher likelihood of response. Citation Format: Joyce F Liu, Su-Chun Cheng, Robert M Wenham, Andrea E Wahner Hendrickson, Deborah K Armstrong, Nancy Chan, David E Cohn, Jung-Min Lee, Richard T Penson, Mihaela Cristea, James Abbruzzese, Koji Matsuo, Alexander B Olawaiye, Jennifer Curtis, Geoffrey Shapiro, Ursula A Matulonis, Elise Kohn, Percy Ivy, Elizabeth M Swisher. Correlation of homologous recombination deficiency (HRD) status by BROCA-HR sequencing with response to combination cediranib and olaparib in relapsed ovarian cancer [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics; 2019 Oct 26-30; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2019;18(12 Suppl):Abstract nr A074. doi:10.1158/1535-7163.TARG-19-A074

BackgroundHomologous recombination deficiency (HRD) affects breast cancer patients. Treatment guided by multigene testing may be particularly beneficial in HRD patients by using platinum-based drugs and poly ADP-ribose polymerase inhibitor (PARPi). However, the optimal method for HRD testing remains undetermined by guidelines or consensus and economic disparities limit the availability of genetic testing. Prioritizing HRD testing by clinical-genomic characteristics is critical for efficient utilization of healthcare resources and improved treatment accuracy.MethodsA total of 93 breast cancer patients who underwent HRD genetic testing were included in the study. According to the machine learning model called genomic scar (GS) HRD was defined as a genomic scar score (GSS) ≥ 50 or with deleterious mutation in the BRCA. Multivariate logistic regression analysis was employed to identify the clinical-pathological factors potentially associated with HRD. Suitable variables were selected to construct a predictive model, and the model’s efficacy was evaluated using the area under the receiver operating characteristic (ROC) curve. Internal validation was performed using bootstrap resampling (500 replicates).ResultsPatients harboring pathogenic mutation in BRCA exhibited higher GSS (99.85 vs 36.90). HRD was not detected in 41.75% of patients, and 34.95% had HRD but no BRCA pathogenic mutations. HRD risk in human epidermal factor growth receptor 2 (HER2) low or positive was significantly lower compared to HER2 negative (OR: 0.390, 95% CI: 0.159–0.959, P = 0.040). High Ki- 67 index was strongly associated with HRD (OR: 28.434, 95% CI: 3.283–246.293, P = 0.002). Significant variations in GSS were observed based on estrogen receptor (ER) and progesterone receptor (PR) status, histological grade, and molecular types. The area under the ROC curve (AUC) of the combined prediction model combining HER2 status and Ki- 67 index was 0.749, and the accuracy of the model was further validated using bootstrap resampling (500 replicates), resulting in an AUC of 0.730, indicating a high predictive accuracy for HRD status.ConclusionsBRCA mutation status did not fully reflect HRD status. Patients with a negative HER2 status and high Ki- 67 index are more likely to exhibit positive results when undergoing HRD genetic testing. The ER, PR, HER- 2 status, Ki- 67 index, molecular typing, and histological grading may have a strong influence on the HRD status.

16P Homologous recombination deficiency in endometrial cancer: Association with clinical and molecular characteristics

Background: Unfavorable vasculogenesis, such as lymphangiogenesis and hemoangiogenesis are known to be pivotal of cancer progression. Our group reported that S1P pathway link chronic inflammation, such as obesity, and cancer progression by inducing lymphangiogenesis and hemoangiogenesis in breast cancer. During the past decade, complex cytokine network interactions between cancer cells and stromal cells have been found to mediate tumor-associated lymphangiogenesis and hemoangiogenesis. Thus, it is crucial to understand cancer biology as a tumor microenvironment, rather than focusing solely on cancer cells. To this end, we hypothesized that amount of intratumoral lymphatic endothelial cells (iLEC) reflect lymphangiogenesis and associate with poor clinical characteristics and outcomes. Method: We analyzed a total of 4145 breast cancer patient’s clinical and transcriptome data from The Cancer Genome Atlas (TCGA), and GSE96058 datasets. Intratumoral lymphatic endothelial cells (iLEC) and all the other immune cells were determined by xCell algorithm, and a median cutoff was adopted to divide high and low groups. Result: First we performed gene set enrichment assay using Hallmark and PID collections from MSigDB and found that high iLEC breast cancer enriched Lymphangiogenesis-related gene sets (lymph angiogenesis, S1P meta, S1P1, and S1P3 pathway) and hemoangiogenesis-related gene sets (Angiogenesis, VEGFR1 pathway, and VEGFR2 pathway) (all FDR<0.25). Microvascular endothelial cells were also predominant in high iLEC tumors (p < 0.001). Even though iLEC was significantly higher in patients with lymph node metastasis consistently in both TCGA and GSE96058 cohorts, there was no survival difference by the amount of iLEC in neither of the cohorts. High iLEC breast cancer also enriched cancer stem cell-related gene sets (notch signaling, hedgehog signaling, epithelial mesenchymal transition, and Wnt β catenin signaling) and metabolic pathways (bile acid metabolism, fatty acid metabolism), and S1P2 pathway (all FDR<0.25). Stromal cells such as Adipocytes, Preadipocytes and Fibroblasts, were also more prominent in the high iLEC group (all p < 0.001). The immune-related gene sets (allograft rejection, IL2 STAT5 pathway, IL6 Jak STAT3 signaling, inflammatory response, TNFα signaling via NFkB, and complement) and cytolytic activity, reflecting anti-tumor immunity, were elevated in high iLEC group of both cohorts (all FDR<0.25, p < 0.001), but anti-cancer immune cells were not significantly infiltrated. Interestingly, cell proliferation-related gene set (E2F target, G2M checkpoint, MYC target v1) and mTORc1 signaling gene set was strongly enriched in low iLEC breast cancer (all FDR<0.25). Furthermore, Ki67 expression and histological grade as well as aggressive cancer subtype were negatively correlated with iLEC fraction (all p<0.002, r <-0.29). Finally, we found that in the TCGA cohort, low iLEC breast cancer was significantly associated with intratumor heterogeneity, silent and nonsilent mutation rate, fraction altered, and homologous recombination deficiency, based on scoring by Thorrson et al (all p < 0.001). Despite the association with high TIL regional fraction and IFN-γ Response, low iLEC breast cancer was infiltrated by pro-cancer immune cells, such as Th2, Treg, and B-cells (all p < 0.001). Conclusion: High iLEC breast cancer reflected lymphangiogenesis and were associated with cancer stem cell-related genes and lymph node metastasis, while low iLEC breast cancers were associated with accelerated cell proliferation and pro-cancer immune cell infiltrations. Unfavorable cancer characteristics on both high and low iLEC groups may explain the lack of difference in patient survival. Citation Format: Rongrong Wu, Masanori Oshi, Mariko Asaoka, Akimitsu Yamada, Yamato Takabe, Li Yan, Itaru Endo, Takashi Ishikawa, Kazuaki Takabe. Intratumoral lymphatic endothelial cell infiltration reflects lymphangiogenesis and lymph node metastasis, but is counterbalanced by immune response and better cancer biology in breast cancer tumor microenvironment [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 P5-06-03.

519Background: HRD status is significantly associated with a higher rate of response to neoadjuvant platinum-based therapy and improved PFS following adjuvant doxorubicin and cyclophosphamide (AC) in TNBC. We assessed the prognostic and predictive role of the HRD assay for platinum and PARP inhibitor response in BrighTNess. Methods: 634 stage II-III TNBC pts were randomized 2:1:1 to: Arm A: Paclitaxel (T) q wk x 12 + carboplatin (P) (AUC 6) q3 wk x 4 + veliparib (TPV) - > AC q2-3 wk x 4; Arm B: T + P + placebo (TP) - > AC; or Arm C: T + dual placebo (T) - > AC. HRD status was defined as HRD+ (HRD score ≥ 42 or a tumor BRCA1/2 mutation) or HRD- (HRD score < 42 and no tumor BRCA1/2 mutation). An exploratory HRD threshold of ≥ 33 vs < 33 was also assessed. Results: HRD status was available for 438 pts. HRD data by arm for pCR for the 42 and 33 cut-offs are shown. Within each arm using the 42 and 33 cut-offs, respectively, ORs for pCR by HRD status (HRD+/HRD-) were 2.85 (p = 0.0005) and 3.10 (p = 0.004) for A...

31P Accurate detection of HRD status in multiple cancer types using somatic mutation pattern

Background: Both HR deficiency and BRCA mutation status predict response to platinum-based therapy and BRCA mutation status predicts docetaxel resistance. However, the association of either biomarker with response to the individual elements of either AC or taxanes (T) is unknown since T is commonly given concomitantly with or after anthracyclines (A). We evaluated the association of HRD and BRCA mutation status with response to neoadjuvant weekly T followed by AC or (F)EC in high-risk breast cancer. Methods: We studied 140 high risk Stage I-III breast cancer patients (pts), enrolled in the breast cancer genome guided therapy study (BEAUTY), obtaining biopsies for DNA/RNA sequencing and MRI imaging to assess response to neoadjuvant weekly T (+trastuzumab+/-pertuzumab for HER2+ disease) followed by AC or (F)EC. Germline BRCA status and HR status of tumor samples (Myriad laboratories) were obtained. HR deficient tumor was defined as HRD score ≥42 or BRCA mutation. MRI response by changes in tumor size after 12 weeks of T was classified by WHO criteria. pCR was defined as ypT0/Tis ypN0. Both MRI response after T and pCR (after T and AC) were examined in terms of germline BRCA mutation (gBRCAmut vs. gBRCAwt) and tumor HR deficiency. Results: Of 140 pts enrolled, 8 withdrew consent and 2 carboplatin treated pts were excluded. Germline data were available for 124/130 pts. 12 patients had BRCA deleterious germline mutations (4 BRCA1, 8 BRCA2). MRI partial (PR)/complete response (CR) rate to T was 47.3% (95% CI: 37.8-57.0%) in the BRCAwt group and 66.7% (95% CI: 34.9-90.1%) in the BRCAmut group. No MRI CR's were observed in BRCA1 mut pts. In contrast, pCR rate was 50% in the 12 gBRCAmut pts (95% CI: 21.1-78.9%) and 31.3% in the 112 gBRCAwt pts (95% CI: 22.8-40.7%). HR deficiency status has thus far been determined for 74 pts: 26 pts have HD deficient tumors: 18 TNBC, 5 Luminal B, 2 ER-/HER2+; and 1 ER+/HER2+. Determination of HR deficiency is ongoing and will be reported for the full cohort in terms of 12 week MRI response to T and pCR to T+AC. HR deficientMolecular Subtypeyes (%)no (%)TBD (%)Luminal A0/112/11 (18.2)9/11 (81.8)Luminal B5/37 (13.5)13/37 (35.1)19/37 (51.3)Luminal NOS0/21/2 (50)1/2 (50)ER+/Her2+1/17 (5.8)14/17 (82.4)2/17 (11.8)ER-/Her2+2/20 (10)11/20 (55)7/20 (35)Triple Negative18/43 (41.9)6/43 (18.6)17/43 (39.5)germline BRCA statusMRI partial response after T (%)MRI complete response after T (%)pCR after T&amp;AC (%)BRCA11/4 (25)0/42/4 (50)BRCA25/8 (62.5)2/8 (25)4/8 (50)BRCAwt35/112 (31.3)18/112 (16.1)35/112 (31.3) Conclusion: In the setting of neoadjuvant weekly T followed by AC, pCR rates were non-significantly higher in pts with BRCA1 mutations. While we observed no overall association between BRCA mutation status and response rates to taxanes; nearly all MRI responses to taxanes (partial and complete) were observed in the BRCA2 group. Prospective studies are needed to validate these findings and to determine whether BRCA status can be used to select therapy. HR deficiency is uncommon in luminal A and HER2+, frequent in TNBC, and the association of HRD with both MRI response to taxanes and pCR will be reported at the meeting. Citation Format: Boughey JC, Kalari KR, Suman VJ, McLaughlin SA, Moreno Aspitia A, Moyer AM, Northfelt DW, Gray RJ, Vedell PT, Tang X, Dockter TJ, Jones KN, Felten SJ, Conners AL, Hart SN, Visscher DW, Wieben ED, Ingle JN, Hartman A-R, Timms K, Elkin E, Jones J, Wang L, Weinshilboum RW, Goetz MP. Role of germline BRCA status and tumor homologous recombination (HR) deficiency in response to neoadjuvant weekly paclitaxel followed by anthracycline-based chemotherapy. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-29.

Homologous recombination deficiency (HRD) is recognized as a pan-cancer predictive biomarker that potentially indicates who could benefit from treatment with PARP inhibitors (PARPi). Despite its clinical significance, HRD testing is highly complex. Here, we investigated in a proof-of-concept study whether Deep Learning (DL) can predict HRD status solely based on routine hematoxylin & eosin (H&E) histology images across nine different cancer types. We developed a deep learning pipeline with attention-weighted multiple instance learning (attMIL) to predict HRD status from histology images. As part of our approach, we calculated a genomic scar HRD score by combining loss of heterozygosity (LOH), telomeric allelic imbalance (TAI), and large-scale state transitions (LST) from whole genome sequencing (WGS) data of n = 5209 patients across two independent cohorts. The model's effectiveness was evaluated using the area under the receiver operating characteristic curve (AUROC), focusing on its accuracy in predicting genomic HRD against a clinically recognized cutoff value. Our study demonstrated the predictability of genomic HRD status in endometrial, pancreatic, and lung cancers reaching cross-validated AUROCs of 0.79, 0.58, and 0.66, respectively. These predictions generalized well to an external cohort, with AUROCs of 0.93, 0.81, and 0.73. Moreover, a breast cancer-trained image-based HRD classifier yielded an AUROC of 0.78 in the internal validation cohort and was able to predict HRD in endometrial, prostate, and pancreatic cancer with AUROCs of 0.87, 0.84, and 0.67, indicating that a shared HRD-like phenotype occurs across these tumor entities. This study establishes that HRD can be directly predicted from H&E slides using attMIL, demonstrating its applicability across nine different tumor types.

Deep learning-based classification and mutation prediction from histopathological images of hepatocellular carcinoma.

Overall, 50% of patients with ovarian cancer harbor a BRCA1/BRCA2 mutation or have tumors characterized by homologous recombination deficiency (HRD). We hypothesized that germline BRCA-positive or somatic BRCA+/HRD+ status would be associated with improved survival compared to BRCA-negative (BRCA-)/HRD- status. We performed a retrospective analysis of patients with advanced high-grade ovarian cancer treated from April 2013 to June 2019 and had known germline BRCA and HRD status. Clinical outcomes were stratified by (1) germline BRCA+ (2) germline BRCA- and somatic BRCA+/HRD+, or (3) BRCA-/HRD- status. Progression free and overall survival were estimated using Kaplan-Meier methods stratified by HRD status and modeled via Cox proportional hazards regression. A total of 187 patients met the inclusion criteria: 55 were BRCA-/HRD-, 106 were germline BRCA+, and 26 were somatic BRCA+/HRD+. HRD- tumors had more non-serous histology (20% vs 6% and 0%, p = .04), were less likely to have complete gross resection (R0) at tumor reductive surgery (60% vs 83% and 77%, p = .02) and required more neoadjuvant chemotherapy cycles (4 vs 3 and 3 cycles, p = .03). Patients with BRCA-/HRD- status had worse progression free survival (14.9 months) than those with germline BRCA+ (23.5 months) or somatic BRCA+/HRD+ (20.2 months, p < .001). Patients with BRCA-/HRD- status also had worse overall survival (42.3 months) than those with germline BRCA+ (68.8 months) or somatic BRCA+/HRD+ (69.2 months). On multi-variate analysis, R0 resection at tumor reductive surgery (hazard ratio [HR] 2.83, 95% confidence interval [CI] 1.58 to 5.06, p < .001), and BRCA+/HRD- status (HR 0.52, 95% CI 0.36 to 0.74, p < .001) were significant factors impacting progression free survival. Tumor reductive surgery (HR 0.15, 95% CI 0.07 to 0.31), and BRCA-/ HRD- (HR 2.25, 95% CI 1.08 to 4.70, p = .03) were significantly associated with overall survival. g/sBRCA+/HRD+ ovarian cancer status is associated with improved survival regardless of primary tumor reductive surgery or neoadjuvant chemotherapy. BRCA-/HRD- status is an adverse prognostic factor for survival.

ObjectivesNeoadjuvant chemotherapy (NACT) has been regarded as a standard treatment for those advanced epithelial ovarian cancer patients with massively disseminated tumors. The homologous recombination deficiency (HRD) status has guiding significance...

Modern management of ovarian cancer (OC) relies on molecular diagnostics, with genetic testing playing a central role in therapeutic decisions. High-grade serous ovarian cancer (HGSOC) is frequently associated with mutations in the BRCA1 and BRCA2 genes, as well as other alterations within the homologous recombination repair (HRR) pathway. The identification of pathogenic variants is critical for selecting patients eligible for treatment with poly (ADP-ribose) polymerase inhibitors (PARPi), which significantly improve progression-free survival, especially in individuals with BRCA mutations and homologous recombination deficiency (HRD). Current guidelines recommend BRCA testing at diagnosis for all patients with HGSOC, followed by HRD testing. Various techniques are used to assess genetic alterations and HRD status. Commercial tests assess mutations in genes in HRR pathways, genomic instability, or HRR functional status to quantify HRD. Despite the availability of these assays, challenges remain regarding test standardisation, predictive accuracy, and cost-effectiveness. Moreover, emerging research highlights the potential for artificial intelligence (AI) to enhance molecular profiling, utilising whole-slide imaging (WSI) and deep learning to predict homologous recombination deficiency (HRD) and other tumour characteristics. The integration of molecular subtypes, as defined by The Cancer Genome Atlas (TCGA), into routine clinical practice holds promise for tailoring therapy beyond BRCA or homologous recombination deficiency (HRD) status. As the field advances, comprehensive genetic testing combined with AI-driven analytics may become the cornerstone of precision oncology in ovarian cancer.

Multi-omic profiling of tumor-infiltrating T cells provides new insights into the differences in the effectiveness of CDK4 inhibitor between human papillomavirus (HPV)-positive and HPV-negative head and neck squamous cell carcinoma patients.