Browse Articles

m-plane GaN split-well direct-phonon terahertz quantum cascade laser

A comprehensive discovery platform for ELOVL1 small-molecule inhibitors targeting very long-chain fatty acid synthesis in adrenoleukodystrophy

Efficacy and safety of antibody-drug conjugates in gastrointestinal malignancies: A systematic review and meta-analysis of phase I-III trials.

e15035 Background: Antibody-drug conjugates (ADCs) represent a promising targeted therapy class for GI malignancies, with emerging activity across HER2, Trop-2, CLDN18.2, and other targets increasingly supporting tumor agnostic approaches. We conducted the first comprehensive systematic review and meta-analysis to synthesize efficacy and safety outcomes of ADCs across all GI malignancies from phase I-III trials. Methods: We conducted a systematic review and meta-analysis of phase I-III trials evaluating ADCs in GI malignancies, searching PubMed, EMBASE, ClinicalTrials.gov, Web of Science, Cochrane Library, and ASCO/ESMO/AACR proceedings through December 2025. Study-level data were pooled using random-effects models (R metafor package) with logit transformation for proportions (objective response rate [ORR], disease control rate [DCR], adverse events [AEs]) and log-scale pooling for medians (PFS, OS, DOR) with variance from 95% CIs, exponentiated to months. Restricted maximum likelihood estimation was employed. Results: Thirty-eight trials (N = 2,893 patients) were included, evaluating HER2-targeted, Trop-2, CLDN18.2, and other ADCs in advanced/pretreated GI cancers. The pooled ORR was 28.0% (95% CI: 23.0-33.7; I² = 87.4%), with a DCR of 74.2% (67.0-80.3; I² = 91%). Median PFS was 4.3 months (95% CI: 3.7-5.0; I² = 88.6%), median OS 9.6 months (8.2-11.3; I² = 84.1%), and median DOR 6.3 months (5.7-6.9; I² = 21.4%). Subgroup analyses showed ORR of 35.4% (29.8-41.5) with HER2-targeted ADCs, 33.4% (19.9-50.2) with CLDN18.2, 4.4% (1.4-12.7) with Trop-2; ORR was 32.7% (25.5-40.9) with topoisomerase I inhibitor payloads and 30.2% (15.2-51.1) with microtubule inhibitors such as MMAE. ORR was 35% in biliary tract cancers and 31.2% in gastric, with 21.4% in colorectal cancers and 21.7% in pancreatic cancers. ORR remained consistent across subsequent lines of therapy, with 27.0% (23.1-31.4; I² = 69.3%) in ≥2nd line and 26.4% (15.61-41.0; I² = 92.8%) in ≥3rd line settings. Grade ≥3 AE rate was 59.6% (54.5-64.5; I² = 79.4%) overall, with key toxicities including anemia (17.2%; 13.0-22.4), neutropenia (14.3%; 7.2-26.6), and ILD/pneumonitis (ILD; any grade 11.9%, ≥3 3.5%). ILD was higher with HER2 ADCs (any 13.1%, ≥3 3.6%), and neutropenia was predominant in pancreatic cohorts (32.8%). Grade ≥3 AEs occurred in 63.5-68.1% with topoisomerase payloads and 55.3% with microtubule payloads. Conclusions: This comprehensive meta-analysis demonstrates meaningful antitumor activity of ADCs in GI malignancies, supporting broad therapeutic potential, including and extending beyond HER2+ gastric cancer, particularly with novel targets like CLDN18.2 and in biliary/gastric subtypes. Safety profiles appear consistent with known ADC toxicities, though topoisomerase payloads and HER2 targeting confer higher risks of cytopenias and ILD, respectively.

SPECTRUM: Phase II study of tislelizumab plus pemetrexed in patients with relapsed or refractory (R/R) primary diffuse large B-cell lymphoma (DLBCL) of the central nervous system (CNS)—The Korean Cancer Study Group LY22-07.

2009 Background: Primary DLBCL of the CNS is a rare extranodal lymphoma confined to the CNS. Despite standard high-dose methotrexate (HD-MTX)-based induction therapy, approximately 25% of patients (pts) relapse after an initial response, and effective salvage options are limited with reported objective response rate (ORR) generally <30% and median progression-free survival (PFS) <3 months. The SPECTRUM trial evaluated efficacy and safety of tislelizumab plus pemetrexed in pts with R/R primary DLBCL of the CNS (NCT05253118). Methods: SPECTRUM is a phase II, multicenter, open-label study conducted in South Korea. Pts with histologically confirmed R/R primary DLBCL of the CNS after HD-MTX-based therapy were eligible. Tislelizumab (200 mg) plus pemetrexed (500 mg/m 2 ) was administered intravenously every 3 weeks until disease progression or unacceptable toxicity. The primary endpoint was investigator-assessed ORR per IPCG response criteria. Secondary endpoints included PFS, duration of response (DoR), overall survival (OS), and safety. Exploratory analyses assessed baseline CSF and plasma MYD88 L265P mutations using droplet digital PCR and genetic subtypes based on LymphGen classification. Results: As of December 16, 2025, 29 pts were enrolled. Median age was 59 years (range, 41–86), 17 (58.6%) were male, and 25 (86.2%) had prior rituximab exposure. Median follow-up was 30.4 months (95% confidence interval [CI], 16.8–not reached [NR]), with a median of 6 treatment cycles (range, 1–55). Among evaluable pts, the ORR was 71.4% (20/28), with complete response (CR) rate of 50.0% (14/28). Median DoR was 14.5 months (95% CI, 2.8–NR). Median PFS and OS were 6.8 (95% CI, 2.1–27.6) and 27.8 months (95% CI, 7.4–NR), respectively. Treatment-emergent adverse events (TEAEs) of any grade occurred in 19 pts (65.5%), most commonly skin rash (51.7%). Grade ≥3 TEAEs occurred in 10 pts (34.5%), including one grade 5 event. Six pts discontinued treatment due to treatment-related AEs, all attributable to pemetrexed. Immune-related AEs occurred in 5 pts (17.2%). Most pts were MCD subtype (84.6% [22/26]), with MYD88 L265P mutations detected in 61.5% (16/26). At baseline, MYD88 L265P detection was higher in CSF than plasma cell-free DNA (cfDNA) (85.0% [17/20] vs 20.0% [4/20]). In one pt with extra-CNS relapse harboring MYD88 L265P mutation and PIM1 – CD274 fusion, MYD88 L265P became undetectable at Weeks 6 and 12, concordant with metabolic CR on PET/CT. Conclusions: Tislelizumab plus pemetrexed met the primary endpoint and demonstrated encouraging outcomes (ORR 71.4% [CR 50.0%]; median PFS and DoR 6.8 and 14.5 months) with manageable safety profiles in pts with R/R primary DLBCL of the CNS. Clinical trial information: NCT05253118 .

Rethinking risk: Disparities and genetic testing outcomes from a novel public-facing cancer prevention program.

e13607 Background: Identifying individuals at high risk for cancer is central to effective prevention and early detection. We evaluated outcomes from a novel public-facing cancer prevention program designed to broaden engagement beyond traditional referral-based models. Methods: The Hennessy Institute, a cancer risk assessment, early detection, and prevention program at Hackensack Meridian Health, launched in July 2024. Data was retrospectively analyzed through January 2026. A multi-faceted outreach strategy guided individuals to an online assessment of personal and family cancer history identifying participants as "high risk" if they met NCCN criteria for genetic testing (HBOC, Lynch, or polyposis), high risk breast screening, or lung cancer screening. Participants interested in learning more were contacted by a trained nurse navigator to review their assessment results. Genetic testing was offered regardless of guideline criteria. Results: Among 4,680 individuals screened, the majority were female (84%). 1,035 reported ethnicity/ race including: 62% White Non-Hispanic, 20% Hispanic, 9% Black, 7% Asian, and 2% other. In total, 53% of individuals were deemed high risk, meeting at least one of the following criteria: genetic risk (45%), high-risk breast (19%), or high-risk lung (8%). Significant ethnic disparities were found in risk status, with a higher proportion of White and Non-Hispanic subjects being classified as high-risk compared to Non-White and Hispanic subjects. Overall, 39% expressed interest in learning more and agreed to be contacted for follow-up; interest was higher among high-risk compared to average-risk individuals (43% vs 34%, p < 0.001), while age was not associated with interest. Among 332 individuals who completed genetic testing, 49 tested positive for pathogenic/ likely pathogenic variants (15%). Mutation positivity rates did not significantly differ between individuals meeting NCCN criteria and those who did not (15% vs 13%, p = 0.59). Conclusions: This real-world analysis from a public cancer prevention program revealed significant ethnic disparities in risk stratification and a greater than expected mutation positivity rate within a general population cohort. These findings suggest restricting genetic testing to guideline-defined populations may miss clinically meaningful cancer risk and highlight the need to evaluate broader genetic screening strategies. Group Analyzed (mean) t Stat P(T≤ t) one-tail P(T≤ t) two-tail High Risk Rate White (57%) vs. High Risk Rate Non-White (42%) ** -4.85664 0.0000007 0.0000014 High Risk Rate Hispanic (44%) vs. High Risk Rate Non-Hispanic (54%)* 2.59568 0.0047870 0.0095740 Interest Rate Average Risk (34%) vs Interest Rate High Risk (43%)** -5.71185 0.0000000 0.0000000 Positivity Average Risk (13%) vs Positivity met NCCN criteria (15%) -0.54341 0.0292575 0.585151 *Significant p < 0.01. **Significant p < 0.001.

National Quality Improvement Initiative to Increase Smoking Cessation Assistance in Commission on Cancer Programs and National Accreditation Program for Breast Centers

PURPOSE Smoking cessation after a cancer diagnosis improves survival, but widespread adoption of evidence-based cessation assistance has not been demonstrated. American College of Surgeons' accredited cancer programs participated in the nationwide Beyond ASK quality improvement (QI) initiative to increase the proportion of currently smoking patients with cancer offered cessation assistance as part of cancer care delivery. METHODS A national QI project was employed between January 2023 and January 2024 following the Plan-Do-Study-Act methodology, and five longitudinal surveys were administered. Participating programs received educational webinars, an online practice change package that contained information about evidence-based smoking assessment and cessation assistance tools, training opportunities, and electronic health record guidance. Primary outcomes included identification of current smoking among patients with newly diagnosed cancer and rate of providing cessation assistance among currently smoking patients. RESULTS A total of 324 programs (164 [50.8%] community programs) enrolled in Beyond ASK. Participation rates were high with 300 (92.6%) programs completing all five surveys. Among 446,015 reported patients newly diagnosed with cancer, 52,794 (11.8%) were identified as currently smoking of which 33,638 (63.7%) received cessation assistance. The mean assist rate increased from 48.0% (95% CI, 43.7 to 52.2) at baseline to 67.5% (95% CI, 63.6 to 71.3) at final. Full adoption was reported by 65.4% of programs. Delivery of cessation assistance increased over time for in-office brief counseling (33.9%-65.8%, P = .0002), in-office behavioral counseling (7.1%-18.5%, P = .02), referral to in-house program (14.5%-27.3%, P = .02), referral to community program (12.1%-29.5%, P = .002), and referral to web-based programs (12.2%-33.9%, P = .0002). CONCLUSION Scaled improvement in smoking cessation assistance across accredited cancer programs is feasible and achievable relatively quickly. Findings provide a framework to guide national adoption for smoking cessation assistance as standard care for all patients with newly diagnosed cancer.

Rezvilutamide (Rez) versus bicalutamide (Bic) plus androgen-deprivation therapy (ADT) in high-volume, metastatic, hormone-sensitive prostate cancer (mHSPC): 5-year outcomes of the phase 3 CHART trial.

5090 Background: In the CHART trial, Rez plus ADT significantly improved radiographic progression-free survival (rPFS) and overall survival (OS) compared with Bic plus ADT in patients (pts) with high-volume mHSPC (Gu et al., Lancet Oncol 2022). Here, we present updated outcomes ~5 years after the enrollment of the last pt. Methods: Pts with high-volume mHSPC and without previous chemotherapy or other localized treatment for PC were eligible to be enrolled. Pts were randomized 1:1 to receive ADT plus either Rez (240 mg) or Bic (50 mg) orally once daily. Final analysis for OS was planned after 325 deaths were reported. Results: Between Jun 28, 2018, and Aug 6, 2020, 654 pts were randomized to receive Rez plus ADT (n=326) or Bic plus ADT (n=328). As of data cutoff (Jun 6, 2025), with a median follow-up of 72.8 months (IQR 69.1-77.4), there were 141 (43.3%) deaths in Rez plus ADT group and 194 (59.1%) in Bic plus ADT group. OS was improved with Rez plus ADT versus Bic plus ADT (median 78.8 months [95% CI 68.3-not reached (NR)] vs 44.8 months [95% CI 37.1-57.4]; HR 0.59 [95% CI 0.47-0.73]; 2-sided p<0.0001). Risk of progression or death in Rez plus ADT group was reduced by 62% compared with Bic plus ADT group (HR 0.38 [95% CI 0.30-0.48]). Other efficacy outcomes also favored Rez plus ADT (Table). Rez plus ADT improved patient-reported quality of life. No new safety signals were observed. Conclusions: Rez plus ADT continued to show meaningful survival benefits with manageable toxicities after long-term follow-up, further supporting Rez plus ADT as frontline treatment for pts with high-volume mHSPC. Clinical trial information: NCT03520478 . Efficacy endpoints. Rez plus ADT (n=326) Bic plus ADT (n=328) HR (95% CI) 2-sided p-value OS, months 78.8 (68.3-NR) 44.8 (37.1-57.4) 0.59 (0.47-0.73) <0.0001 rPFS per investigator, months 81.0 (62.6-NR) 18.5 (14.8-25.7) 0.38 (0.30-0.48) <0.0001 rPFS per investigator (sensitivity analysis # ), months 58.9 (50.10-NE) 18.6 (16.30-22.90) 0.38 (0.31-0.47) <0.0001 Time to PSA progression, months NR (NR-NR) 11.0 (9.2-12.9) 0.21 (0.17-0.27) <0.0001 Time to next skeletal-related event, months 78.8 (58.9-NR) 37.7 (34.3-48.2) 0.63 (0.51-0.78) <0.0001 Time to initiation of new anti-prostate cancer therapy, months 62.4 (46.4-75.6) 15.1 (13.6-18.1) 0.32 (0.26-0.39) <0.0001 Data are median (95% CI) unless otherwise indicated. # New anti-cancer therapy and missing tumor assessments were not considered. PSA, prostate specific antigen.

Clinical characterization and survival of patients (pts) with <i>PIK3CA</i> -mutated metastatic colorectal cancer (mCRC): Individual patient data (IPD) meta-analysis of four randomized trials.

e15541 Background: We aimed to characterize pts with PIK3CA MUT mCRC and to estimate survival within an IPD meta-analysis of the randomized trials FIRE-1 (FUFIRI vs. mIROX), CIOX (CAPIRI or CAPOX plus Cetuximab), FIRE-3 (Cetuximab or Bevacizumab plus FOLFIRI) and PanaMa (FU/FA ±.Panitumumab after FOLFOX+Panitumumab). Methods: Kaplan-Meier estimated the progression-free (PFS) and overall survival (OS). Hazard ratios (HR) including the 95% confidence interval (CI) and interaction of PIK3CA status and treatment (anti-EGFR antibody vs. no anti-EGFR antibody) were assessed by adjusted multivariable Cox proportional hazard models (study, age, ECOG, primary tumor location). Discrete variables were compared by Fisher’s exact test. Results: In 107 of 920 tumors (11.6%), a PIK3CA MUT was detected. Concurrent RAS and BRAF MUT were found in n = 31 (29.0%) and n = 10 tumors (9.3%). A trend towards higher age ( P = 0.12), lower performance status ( P = 0.09) and right-sided primary tumors ( P = 0.14) was observed in PIK3CA MUT mCRC pts, in addition to atypical metastatic sites ( P = 0.047; other than liver, lung, nodes or peritoneum). Table 1 shows the median PFS and OS according to the PIK3CA status. In multivariate analyses, PIK3CA was not prognostically relevant for PFS ( P = 0.24), while a significant effect was observed for OS ( P &lt; 0.001). The use vs. omission of anti-EGFR antibodies was irrelevant for PFS, regardless of PIK3CA status (all-WT: HR 1.03, 95% CI 0.81 – 1.32, P = 0.80; PIK3CA MUT: HR 1.34, 95% CI 0.80 – 2.24, P = 0.27; interaction P = 0.41). OS, however, was significantly longer if anti-EGFR antibodies were applied in all-WT tumors (all-WT: HR 0.68, 95% CI 0.53 – 0.88, P = 0.003), while this was not the case for PIK3CA MUT tumors (HR 0.91, 95% CI 0.53 – 1.55, P = 0.73; interaction P = 0.067). Conclusions: This retrospective IPD meta-analysis indicates that PIK3CA mutation might have a prognostic relevance for mCRC and predictive implications for anti-EGFR treatment efficacy. Median PFS and OS according to PIK3CA status in four randomized trials (FIRE1, CIOX, FIRE3. PanaMa). Variable PFS(in months) OS(in months) RAS/BRAF WTPIK3CA WTn=493 9.4(8.8-10.0) 27.2(24.9-29.5) RAS/BRAF WTPIK3CA MUTn=66 8.8(7.6-10.0) 28.0(24.0-32.0) RAS MUTPIK3CA WTn=256 8.7(7.8-9.7) 21.1(18.0-24.2) RAS MUTPIK3CA MUTn=31 8.1(6.3-9.8) 18.5(12.0-25.0) BRAF MUTPIK3CA WTn=64 7.0(4.4-9.6) 15.9(11.7-20.1) BRAF MUTPIK3CA MUTn=10 5.8(4.0-7.5) 17.1(0.0-34.3) P 0.20 &lt;0.001

Comparison of a highly sensitive, tumor-informed Illumina whole genome sequencing research assay and a tumor-informed bespoke whole exome sequencing assay for molecular residual disease detection in CheckMate-77T (NCT 04025879).

8591 Background: In CheckMate 77T (NCT 04025879), nivolumab showed clinically meaningful improvements in event free survival and pathologic complete response vs placebo in 461 patients with resectable (stage IIA-IIIB) NSCLC. Previously, circulating tumor DNA (ctDNA) clearance rates of 66% in NIVO and 38% in PBO of the neoadjuvant period (C1D1 to definitive surgery) had been reported from 130 patients by Provencio et. al. (LBA50, ESMO 2024). In this study, we aim to extend on the previous exploratory study by using a highly sensitive whole genome sequencing (WGS) minimal residual disease (MRD) assay to detect ctDNA clearance rates and compare to those measured using a tumor-informed whole exome sequencing (WES) MRD assay. Methods: Residual samples (tumor FFPE tissue/isolated DNA, normal, cfDNA) from 61 of the 130 participants previously reported comprised this study. The Illumina WGS MRD assay employs WGS of tumor and matched normal to generate tumor fingerprints which were bioinformatically monitored in plasma using low-pass WGS (lp-WGS). All samples were prepared using Illumina’s WGS Oncology Prep (research use only) and analyzed with DRAGEN MRD pipeline. A comparison of the operational complexity and turn-around-time (TAT) between tumor-informed WES MRD assays and the Illumina WGS MRD assay was performed. Results: Detection rates at C1D1 were greater in this study than previously reported, 91% for WGS vs 86% for WES, supporting the hypothesis that WGS MRD is more sensitive than WES MRD assays. Using the WES MRD results as reference, the Illumina WGS assay demonstrated a high concordance with positive percent agreement of 100%, negative percent agreement of 63% and overall percent agreement of 95%. Clearance rates of ctDNA in both study arms were calculated, summarized and compared to previous results by Provencio et. al. Compared to tumor-informed MRD assays, the Illumina WGS MRD assay significantly reduced operational complexity and TAT, as it eliminated the need for bespoke panel design, including analysis, manufacturing and assay validation. The Illumina Oncology WGS Prep uses a single streamlined workflow for all MRD sample types. The lpWGS testing of plasma cfDNA required only 2-5 ng cfDNA which can be obtained from a single tube of whole blood compared to 2 tubes of whole blood as required for many tumor-informed MRD assays. The total TAT of the WGS MRD workflow, including both tissue and plasma testing, was 1-2 weeks, substantially shorter than the 4-6 weeks required for most tumor-informed MRD assays. Conclusions: This study found the Illumina WGS MRD research assay to be highly concordant with previous tumor-informed WES MRD results for MRD positive results, with improved sensitivity, reduced TAT, lower cfDNA input requirements and decreased operational complexity. Clinical trial information: NCT04025879 .

Nodulomics: Histopathologic and genomic heterogeneity among multiple malignant pulmonary nodules sampled by robotic-assisted bronchoscopy.

8049 Background: Multiple pulmonary nodules (MPN) pose significant challenges for accurate staging of non-small cell lung cancer (NSCLC). As next generation sequencing (NGS) technologies have become more widely available on smaller amounts of tissue, the prevalence and clinical implications of heterogeneity between MPNs are poorly understood. This study retrospectively evaluates the histopathological and NGS findings after MPN biopsy in a single robotic-assisted bronchoscopy (RaB) procedure. Methods: 191 patients were retrospectively enrolled at 6 geographically diverse centers in the US. All patients had 2+ pulmonary nodules biopsied and staging EBUS in the same procedure. An expert molecular pathologist reviewed NGS data between nodule pairs to determine if they were discordant, concordant, or indeterminate. Results: Of 191 patients, 85 patients had 2 malignant nodules with 15 of those demonstrating histopathologic heterogeneity (e.g., one adenocarcinoma, one squamous cell carcinoma). 26 patients with 2 malignant nodules of similar histopathology demonstrated molecular heterogeneity between nodules. These cases were reviewed by an expert molecular pathologist who determined the following breakdown: -15 discordant pairs (57.7%) -3 concordant pairs (11.5) -8 indeterminate pairs (30.8%) Breakdown by nodule density: -Discordant: 7 same density, 8 different density -Concordant: 3 same density -Indeterminate: 5 same density, 3 different density Breakdown by nodule laterality: -Discordant: 7 ipsilateral, 8 contralateral -Concordant: 2 ipsilateral, 1 contralateral -Indeterminate: 4 ipsilateral, 4 contralateral Staging EBUS was performed in all 26 cases with malignancy identified in 2 cases. NGS sufficiency for NSCLC from three sites with standardized NGS reporting was as follows: 1 st nodule biopsied: 42/54 (77.8%), 2 nd nodule biopsied: 37/51 (72.5%), 3 rd nodule biopsied: 2/4 (50%). In total, 90/109 (82.6%) NSCLC samples had adequate tissue for NGS. In 85 cases with 2 malignant nodules, 35 (41.2%) had all nodules Stage IIB or lower. Notably, 25 (29.4%) of the cases did not have NGS available on one or more of the nodules, thus precluding staging. Conclusions: This study highlights the frequency and clinical relevance of histopathologic and molecular heterogeneity among synchronous MPNs. Importantly, heterogeneity was not limited to contralateral or anatomically distant lesions, challenging the assumption that spatial proximity implies clonal relatedness. Further, this study shows that relying solely on density or radiographic features of a nodule is an incomplete discriminator between separate primary lung cancers and intrapulmonary metastasis. Our findings support the need for an integrated diagnostic paradigm in which sampling of multiple lesions is used to inform clonality, refine staging, and guide precision therapy.

Scalable knowledge distillation for pancreatic cancer diagnosis: Real-time ResNet-based stage stratification with global deployment feasibility.

e16412 Background: Pancreatic cancer remains one of the deadliest cancers, with a 5-year survival below 12%. Early-stage disease is often misdiagnosed as benign lesions or chronic pancreatitis, delaying treatment. Deep learning models show high diagnostic accuracy on cross-sectional imaging, but infrastructure and computational demands limit clinical use. Knowledge distillation can maintain performance while enabling real-time deployment. We evaluated whether a distilled ResNet18 could retain ResNet152-level accuracy for stage stratification and differential diagnosis while supporting global scalability. Methods: We analyzed 2,840 contrast-enhanced CT scans from multiple countries: pancreatic ductal adenocarcinoma (n = 1,400; 700 resectable/borderline-resectable, 700 locally advanced/metastatic), chronic pancreatitis (n = 900), and benign lesions (n = 540). Ground truth was established via multidisciplinary consensus using histopathology, surgical findings, and longitudinal follow-up. A ResNet152 mentor network (60.2M parameters; 28.6 GFLOPs) served as reference. A ResNet18 mentee (11.7M parameters; 1.8 GFLOPs) was trained via temperature-scaled knowledge distillation with regularized cross-entropy loss. Accuracy, sensitivity, specificity, F1-score, and AUROC were measured. Radiologists and oncologists across six continents evaluated clinical utility and deployment feasibility. Results: The distilled ResNet18 achieved 96.6% accuracy for stage stratification (sensitivity 97.4%, specificity 95.8%, AUROC 0.989). Differential diagnosis accuracy was 94.8% for pancreatic cancer vs chronic pancreatitis and 95.6% vs benign lesions. External validation ranged from 93% to 97% across institutions. Mean inference time was 41 ms per image vs 335 ms for ResNet152, enabling real-time integration into CT workflows. Overall, 93.9% of clinicians rated it clinically valuable for triage and staging support. Conclusions: Knowledge distillation allows a lightweight ResNet18 to achieve near–ResNet152 performance while reducing computational cost by over 90%. This approach addresses key translational barriers in pancreatic cancer imaging and supports real-time, globally scalable clinical decision support. Prospective studies will assess its impact on diagnostic delay, surgical referral, and treatment allocation.

MRI-derived tumor burden and growth rate as translational biomarkers: Head-to-head comparison with RECIST and ctDNA in tebentafusp-treated metastatic uveal melanoma.

e21584 Background: Tebentafusp improves survival in metastatic uveal melanoma (mUM), yet RECIST v1.1 responses are uncommon and may not reflect clinical benefit. We evaluated whether early, on-treatment liver MRI volumetric and kinetic biomarkers better predict overall survival (OS) than baseline imaging, ctDNA dynamics, and RECIST. Methods: We retrospectively included 88 consecutive HLA-A*02:01+ mUM patients treated with tebentafusp (2018–2022) with baseline and month-3 (M3 ±1 month) liver MRI. All measurable liver metastases (≥ 5 mm) were segmented to derive total tumor volume at baseline (TTV 1 ) and month 3 (TTV 2 ). Tumor growth rate between scans was computed assuming exponential growth (TGR 3m , %/month). Plasma ctDNA (ddPCR; GNAQ/GNA11/SF3B1) was assessed at baseline and M3 when available (n=68). Associations with OS were tested using Cox models; prognostic discrimination was evaluated using Harrell’s C-index with bootstrap resampling. A three-tier MRI risk score combined median cutoffs for TTV 2 and TGR 3m . Results: Median age was 59 years; median follow-up was 43.5 months. Median TTV 1 was 8.4 cm³ (IQR 1.7–83.8), TTV 2 14.6 cm³ (IQR 2.4–124.6), and TGR 3m 17.5%/month (IQR 5.2–31.7). Inter-reader reproducibility was high (ICC: 0.84 for TTV1, 0.88 for TTV2, 0.90 for TGR 3m ). In multivariable analysis, higher TTV 2 and TGR 3m were independently associated with shorter OS (HR per doubling of TTV₂ = 1.25 ; 95% CI 1.15–1.35; HR per +10%/month TGR 3m , 1.17; 95% CI 1.07–1.28; both P &lt; 0.001), while baseline tumor burden and RECIST response were not retained. The MRI risk score separated three OS strata (log-rank P &lt; 0.001) with median OS not reached (low risk), 34.6 months (intermediate), and 11.3 months (high risk). Prognostic performance favored the MRI score (C-index 0.79; 95% CI 0.71–0.86), outperforming ctDNA-based models (C-index 0.73; P = 0.01) and RECIST response (C-index 0.65; P &lt; 0.001). Conclusions: Early, on-treatment liver MRI biomarkers, month-3 tumor burden (TTV 2 ) and growth rate (TGR 3m ), provide strong, independent prognostic information under tebentafusp and outperform ctDNA kinetics and RECIST. These metrics support MRI-based risk stratification at 3 months as a practical decision tool to guide monitoring and therapeutic adaptation in mUM.

A matched-adjusted indirect comparison (MAIC) of nogapendekin alfa inbakicept-pmln plus bacillus Calmette–Guérin (NAI+BCG) and pembrolizumab in patients with BCG-unresponsive non–muscle invasive bladder cancer (NMIBC) with carcinoma in situ (CIS) ± papillary disease.

e16619 Background: Several FDA approved treatments offer bladder-sparing options for patients with BCG-unresponsive NMIBC CIS ± papillary disease. In the absence of head-to-head trials, indirect treatment comparison (ITC) methodologies can provide robust assessments between approved treatments. An unanchored, matched-adjusted indirect comparison (MAIC) was used to compare the efficacy of nogapendekin alfa inbakicept-pmln + BCG (NAI+BCG) versus pembrolizumab (PEMBRO). Methods: Feasibility assessments between the phase 3 NAI+BCG study (QUILT-3.032; NCT03022825) and a phase 3 PEMBRO study (KEYNOTE-057; NCT0262596) trials confirmed suitability for ITC on efficacy endpoints. Patient-level data from QUILT-3.032 were matched in an MAIC or weighted in a simulated treatment comparison (STC) against aggregate KEYNOTE-057 data. Endpoints included complete response (CR) at 12-months, duration of response (DOR), progression-free survival (PFS), and overall survival (OS). Restricted mean survival time analysis was used for time-based treatment effect for PFS and OS. Five mutually reported baseline variables (age ≥65, sex, ECOG status, race, stage) were included to minimize bias. Differences in administration limited modeling treatment-related adverse events (TRAEs). Effective sample size and E-values assessed weighting adequacy and robustness to unmeasured confounding in modeling. Results: The weighted MAIC analysis shows that NAI+BCG achieved higher CR at 12 months compared to PEMBRO (47.3% vs 18.8%; odds ratio [OR] 3.88 [95% CI: 1.77-8.50]). The adjusted STC base case found that NAI+BCG patients had a longer DOR compared to PEMBRO, with a median difference of 10.65 months (26.85 vs 16.20 [95% CI: 3.70-8.42]) with sensitivity analysis demonstrating similar outcomes. At 43-month cutoff, the median PFS favored NAI+BCG (median not reached) compared to 39.9 months for PEMBRO (hazard ratio: 0.45 [95% CI: 0.17, 1.20]). No significant difference was found in OS. Odds of non-bladder related grade 3-4 TRAEs favored NAI+BCG vs PEMBRO (OR: 0.33 [95% CI: 0.00, 4.09]). E-values supported model robustness to unmeasured confounding. Conclusions: In this MAIC, NAI+BCG provided a higher CR rate and longer DOR compared to PEMBRO for BCG-unresponsive NMIBC CIS patients. The results provide novel comparative effectiveness data that may be helpful in shared clinical decision-making. Findings should be interpreted cautiously due to inherent limitations of unanchored MAICs.

A randomized controlled pilot study of aromatherapy to reduce anxiety and pain during bone marrow aspiration and biopsy.

12140 Background: Bone marrow aspiration and biopsy (BMAB) is associated with substantial anxiety and pain among patients with hematologic malignancies. Nonpharmacologic strategies to improve procedural tolerance are needed. Aromatherapy with lavender essential oil and its primary active constituent, linalool, has demonstrated anxiolytic effects in preclinical and peri-procedural settings but has not been rigorously evaluated during BMAB. Methods: We conducted a single-center, randomized pilot study of adult patients with cancer receiving a BMAB. Subjects were randomized (1:1:1) to lavender essential oil (≥25% linalool), pure linalool oil (≥97%), or placebo (jojoba oil). Aromatherapy was administered via inhalation prior to and throughout the procedure. Primary outcomes included feasibility (&gt;70% completion) and patient satisfaction. Secondary outcomes included anxiety and pain assessed by visual analog scales (VAS) before, during, and after BMAB. Exploratory physiologic outcomes included bispectral index (BIS), an electroencephalogram-based measure of cortical arousal. Results: Forty-five patients were enrolled, with 43 (96%) completing the study, meeting feasibility criteria. Most participants (88%) agreed or strongly agreed that aromatherapy was relaxing and enjoyable. All groups demonstrated significant within-group reductions in VAS-anxiety from pre- to post-procedure, with the largest reduction observed in the linalool arm (−3.1±0.70 (mean±SE); p&lt;0.01), compared with lavender (−1.8±0.67; p=0.02) and placebo (−2.4±0.61; p&lt;0.01). During BMAB, anxiety reduction favored the lavender and linalool arms, with a trend toward between-group significance (p=0.053). Pain increased during BMAB across all groups; however, the linalool arm demonstrated the smallest increase (2.6±0.69; p&lt;0.01)) compared with lavender (3.8±0.93; p&lt;0.01)) and placebo (4.0±1.3; p&lt;0.01). Post-BMAB, between group analysis trended toward significantly less pain with lavender and linalool compared to placebo (p=0.057). Exploratory within-group BIS scores analyses demonstrated significant reductions in average BIS after aromatherapy treatment prior to BMAB with lavender showing the greatest reductions (-6.0±2.0; p=0.01), as compared to linalool (-3.6±1.6; p=0.047) and placebo (-1.6±1.3, p=0.27). No serious adverse events occurred. Conclusions: Aromatherapy using lavender and linalool during BMAB is feasible, safe, and potentially associated with meaningful reductions in procedural anxiety and pain. These preliminary findings support a larger randomized control trial to confirm efficacy and explore integration into routine supportive oncology care. Clinical trial information: NCT06581211 .

Prognostic impact of cancer stem cell markers on survival outcomes in solid tumors: A systematic review and meta-analysis.

e15187 Background: Cancer stem cell (CSC) markers have been implicated in tumor aggressiveness and treatment resistance. Previous evidence suggests that high tumor expression of these CSC markers is associated with worse patient outcomes. However, their overall prognostic value for survival across solid tumors remains to be quantified. We aimed to systematically evaluate whether high expression of CD133, ALDH1, CD44, or SOX2 in solid tumors predicts poorer overall survival (OS). Methods: We performed a comprehensive literature search (Medline, Embase, Scopus, and Cochrane Library) for studies up to January 2026 that assessed immunohistochemical overexpression of CD133, ALDH1, CD44, or SOX2 in solid tumors and reported survival outcomes. Eligible studies were pooled using random-effects meta-analysis. Hazard ratios (HRs) or risk ratios for OS with 95% confidence intervals (CIs) were extracted or calculated. Heterogeneity and publication bias were assessed, and subgroup analyses were conducted by cancer type and marker. Results: Across included studies, high expression of each CSC marker showed a significant adverse impact on OS. Patients whose tumors overexpressed CD133 had substantially worse survival; for example, in non-small cell lung cancer high CD133 tripled the risk of 5-year mortality (RR ≈ 3.19) compared to low expression, and in head-neck cancers it was associated with ~2-fold higher hazard of death (HR ≈ 2.33). Overexpression of ALDH1 was likewise linked to poor outcomes, in colorectal cancer, high ALDH1 was associated with less than half the 5-year survival rate of low expressors (OR = 0.42, 95% CI 0.26–0.68). CD44 overexpression correlated with shorter survival as well, though effect sizes were more moderate (HR = 1.32, 95% CI 1.08–1.61 in colorectal cancer). Notably, CD44 positivity was still consistently indicative of worse prognosis across multiple tumor types (such as hepatocellular carcinoma and lung cancer). SOX2 overexpression also emerged as a significant negative prognostic factor: patients with SOX2-high tumors had markedly reduced survival (pooled HR ≈ 1.65, 95% CI 1.34–2.04) compared to SOX2-low patients. These associations held true in most subgroup analyses, although the magnitude of risk varied by cancer type. Conclusions: High tumor expression of CSC markers is significantly associated with worse overall survival in patients with solid tumors. This meta-analysis provides quantitative evidence that overexpression of these stemness-related markers portends poorer prognosis. Clinically, assessment of CSC marker status may help identify high-risk patients who could benefit from more aggressive therapy or novel treatments targeting CSCs. These findings underscore the prognostic value of CSC markers and support their further evaluation as biomarkers for risk stratification and as potential therapeutic targets in solid malignancies.

Autologous natural killer cell infusion as consolidation therapy after first-line chemoradiotherapy for limited-stage small-cell lung cancer: A randomized, controlled, open-label, single-center phase II clinical study.

8013 Background: Small-cell lung cancer (SCLC) grows rapidly, is aggressive, has a poor prognosis, and tends to recur after treatment. Autologous cellular immunotherapy (CIT) has demonstrated good safety and efficacy across various tumors; however, there are no prospective studies utilizing autologous natural killer (NK) cells for consolidation therapy following first-line chemoradiotherapy for SCLC. Methods: This study aimed to evaluate the safety and efficacy of autologous NK cell infusion (administered every 2 weeks for a total of six courses) as consolidation therapy after first-line standard treatment for limited-stage SCLC, compared to routine follow-up in a randomized, controlled, open-label, single-center phase II clinical trial. The primary endpoint was progression-free survival (PFS, assessed according to Response Evaluation Criteria in Solid Tumors version 1.1), while secondary endpoints included overall survival (OS), the 12- and 24-month PFS rate, the 24- and 36-month OS rate and safety. Results: Forty-three patients with limited-stage SCLC were included in the final analysis, comprising 21 patients in the treatment group who received autologous NK cell infusion after chemoradiotherapy and 22 patients in the control group who underwent routine follow-up. At 6 months, the response to initial chemoradiotherapy was maintained in 27.3% (6/22) of the control patients and 57.1% (12/21) of the treatment patients. The objective response rate (ORR) and PFS rates at 12 months for controls compared to treatment patients were 0% versus 19.0% (4/21) (P &lt; 0.05) and 4.5% (1/22) versus 42.9% (9/21) (P &lt; 0.01), respectively. Compared to the control group, the autologous NK cell consolidation group achieved significantly longer PFS (median 6.5 vs. 11.92 months; hazard ratio [HR] 0.38, 95% confidence intervals [CI] 0.18 to 0.79; P = 0.01) and OS (median 15.6 vs. 27.13 months; HR 0.41, 95% CI: 0.19 to 0.87, P = 0.02), with a particularly pronounced PFS benefit when calculated from the end of chemoradiotherapy (median 8.1 vs.16.3 months; HR 0.35, 95% CI 0.17 to 0.72, P = 0.01). Regarding safety, the incidence of all adverse events (AEs) due to any cause was 50.0% (11/22) in the observation group and 47.6% (10/21) in the treatment group, with most being grade 1–2 and considered unrelated to NK cell infusion. NK cell treatment exhibited good overall safety and tolerability. Additionally, we characterized changes in peripheral blood cell subsets and metabolites before and after treatment, as well as the tumor immune microenvironment characteristics in patients from the treatment group. Conclusions: Autologous NK cell infusion as consolidation therapy after first-line chemoradiotherapy for SCLC yielded promising preliminary PFS and OS results, with a well-tolerated safety profile. Clinical trial information: NCT03410368 .

BNT324-03: Phase 3, randomized, open-label trial of BNT324/DB-1311, a B7H3 ADC, versus docetaxel in patients with taxane-naïve metastatic castration-resistant prostate cancer (mCRPC).

TPS5137 Background: B7H3 is highly expressed in various human cancers, including mCRPC, and overexpression is associated with a poor patient prognosis. BNT324/DB-1311 is an investigational B7H3 ADC comprised of a topoisomerase-I-inhibitor-based payload with a cleavable linker and a drug-to-antibody ratio of ~6 that received FDA Fast-Track Designation for previously treated CRPC. A phase 1/2 study (DB-1311-O-1001, NCT05914116) is currently enrolling several prostate cancer cohorts: post Lutetium-177 [¹⁷⁷Lu]-PSMA-617 radioligand (RLT) (Lu 177-RLT) mCRPC, taxane-naïve mCPRC, and castration-sensitive prostate cancer (CSPC) with suboptimal PSA response, with patients receiving BNT324/DB-1311 alone or with abiraterone or enzalutamide. This study has reported encouraging activity (median radiographic PFS 11.3 months) and a manageable safety profile in heavily pretreated mCRPC (Parsonson A, ASCO 2025 #5015), and in Lu 177-RLT pretreated mCRPC (Stein M, ASCO GU 2026 #176). The treatment landscape of mCRPC has been evolving. As first use of an androgen receptor pathway inhibitor (ARPI) has shifted to the mCSPC setting, more patients will receive docetaxel in early-line mCRPC. The use of Lu 177-RLT in taxane-naïve mCRPC has been increasing; however, outcomes with subsequent docetaxel appear suboptimal. Docetaxel remains the first-line treatment option for symptomatic or rapidly progressive mCRPC, but outcomes are poor. Overall, there is a high unmet need for novel effective and better tolerated treatment options in the post-ARPI and emerging post-RLT settings. Methods: BNT324-03 is an open-label, randomized, Phase 3 trial (NCT07365995) designed to determine the efficacy and safety of BNT324 compared with docetaxel in patients ( &gt; 18 years, ECOG PS 0–1) with mCRPC previously treated with 1 or 2 prior ARPIs. Patients may have received prior Lu 177-RLT but must not have received prior taxane for mCRPC (allowed in mCSPC if ≥6 months since completion without disease progression). Prior treatment with B7H3 targeted therapy is not allowed. Approximately 736 patients will be randomized 1:1 to receive either BNT324 or docetaxel (+ prednisone/prednisolone). Unless there is unacceptable toxicity or withdrawal of consent, BNT324 will be administered until PCWG3-modified RECIST v1.1-defined progressive disease and docetaxel will be administered for up to 10 cycles. Randomization will be stratified by prior Lu 177-RLT, prior taxane use, and site of metastasis. The dual primary endpoints are to assess the efficacy of BNT324 in terms of radiographic PFS and OS when compared with docetaxel. Secondary endpoints include time to first subsequent therapy, ORR and DOR by BICR, time to pain progression, time to first symptomatic skeletal-related event, PSA response and time to PSA progression, and safety. Enrollment is ongoing globally. Clinical trial information: NCT07365995 .

Impact of pre-existing autoimmune disease on clinical outcomes in patients treated with bispecific monoclonal antibodies for B-cell lymphomas.

e19008 Background: Bispecific antibodies (BsAbs), including glofitamab, epcoritamab, and mosunetuzumab, are FDA-approved for relapsed/refractory B-cell lymphomas. Patients with pre-existing autoimmune disease (AD) were excluded from pivotal trials, resulting in limited real-world data on outcomes in this population. We evaluated the impact of pre-existing AD on clinical outcomes in patients with follicular lymphoma (FL) and diffuse large B-cell lymphoma (DLBCL) treated with BsAbs. Methods: We performed a multicenter retrospective cohort study using the TriNetX database. Two cohorts were identified: patients with FL or DLBCL and a documented pre-existing AD treated with BsAbs, and a comparator cohort without AD treated with BsAbs. Patients were matched 1:1 using propensity score matching based on age, sex, race, comorbidities, prior CAR T-cell therapy, and baseline laboratory values. Outcomes included 1-year and 3-year overall survival (OS); cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS) within 60 days; and infections, neutropenia, and ICU admissions within 90 days of BsAb initiation. Survival was assessed using Kaplan–Meier analysis, and outcomes were summarized using hazard ratios (HR), risk ratios (RR), and 95% confidence intervals (CI). Results: In the unmatched cohorts, 257 patients had a pre-existing AD and 631 patients comprised the non-AD control group. Following propensity score matching, 228 patients were included in each cohort. There was no statistically significant difference in one-year or three-year OS between patients with pre-existing AD and those without AD (1-year HR = 0.85 [95% CI, 0.62–1.18]; 3-year HR = 0.93 [95% CI, 0.69–1.26]). Rates of CRS (RR = 1.13 [95% CI, 0.87–1.46]) and ICANS (RR = 1.00 [95% CI, 0.64–1.58]) within 60 days were similar between cohorts. Infection rates (RR = 1.14 [95% CI, 0.91–1.43]) and ICU admissions (RR = 1.09 [95% CI, 0.71–1.65]) within 90 days did not significantly differ. Patients with pre-existing AD had a higher risk of neutropenia within 90 days compared with those without AD (RR = 1.26 [95% CI, 1.04–1.52]). Conclusions: In this large real-world analysis pre-existing autoimmune disease was not associated with inferior overall survival or increased rates of CRS, ICANS, infection, or ICU admission in patients with relapsed/refractory FL or DLBCL. Pre-existing AD was associated with a higher risk of neutropenia, underscoring the importance of close hematologic monitoring. Outcome Pre-Existing AD(n= 228) No-AD(n=228) HR/RR [95% CI] 1-year OS 63.4% 57.9% 0.85 [0.62-1.18] 3-year OS 44.5% 45.6% 0.93 [0.69-1.26] CRS (≤60 days) 35.1% 31.1% 1.13 [0.87-1.46] ICANS (≤60 days) 14.0% 14.0% 1.00 [0.64-1.58] Infection (≤90 days) 42.5% 37.3% 1.14 [0.91-1.43] Neutropenia (≤90 days) 53.9% 43.0 1.26 [1.04-1.52] ICU Admission (≤90 days) 16.7% 15.4% 1.09 [0.71-1.65]

Efficacy and safety outcomes of biweekly compared with standard triweekly gemcitabine/nab-paclitaxel in advanced pancreatic cancer: A systematic review and meta-analysis.

e16405 Background: Pancreatic cancer remains one of the most lethal malignancies, with a rising incidence and persistently poor survival outcomes. Gemcitabine plus nab-paclitaxel is an established standard regimen for advanced disease, however, the conventional triweekly schedule is frequently associated with significant toxicity. Biweekly dosing has been proposed to enhance tolerability while maintaining efficacy, although evidence remains limited and heterogeneous. The objective was to evaluate the efficacy and safety of biweekly versus standard gemcitabine/nab-paclitaxel in patients with advanced pancreatic cancer. Methods: A systematic search of PubMed, Embase, and the Cochrane Library was conducted to identify studies comparing Biweekly versus Standard Triweekly: gemcitabine / nab-paclitaxel in patients with advanced pancreatic cancer. The primary outcomes of interest included partial response, and treatment-related toxicities, specifically anemia, neutropenia, and use of growth factors. Statistical analyses were performed using Review Manager. Heterogeneity was assessed using the I 2 statistic, and risk of bias was evaluated using the ROBINS-I tool. Results: A total of 1357 patients from eight observational studies were included. Comparative analysis showed no significant difference in partial response between biweekly and standard triweekly gemcitabine/nab-paclitaxel in patients with advanced pancreatic cancer with a risk ratio (RR) of 0.77 (95% CI 0.57,1.04, p = 0.09). Similarly, no significant differences were observed between regimens in the incidence of anemia (RR = 1.09; 95% CI 0.74–1.59; p = 0.66) or neutropenia (RR = 1.09; 95% CI 0.86–1.37; p = 0.48). however, growth factor use was significantly lower with biweekly dosing (RR = 0.71; 95% CI 0.58–0.87; p = 0.0009). Overall survival was comparable between groups RR = 1.02 (95% CI 0.71, 1.45 p = 0.93). Conclusions: Administration of gemcitabine/nab-paclitaxel biweekly yielded clinical outcomes and safety profiles comparable to the conventional regimen, while significantly reducing the need for growth factors as supportive care. These data suggest that a biweekly regimen could be an effective alternative; however, further prospective randomized studies are necessary to verify these findings and better understand their clinical application.

STRATOS-P clinical model for prognosis and selection of patients with metastatic hormone sensitive prostate cancer (mHSPC) for intermittent therapy.

5106 Background: Current trials use prostate-specific antigen (PSA) response of ≤0.2 ng/mL at 6-12 months as a prognostic marker and identify patients for intermittent therapy. Clinico-genomic assessment can increase prognostic accuracy and patient selection for intermittent therapy, which would improve quality of life and decrease adverse events associated with mHSPC combination therapies without impacting survival. Methods: Retrospective study of veterans diagnosed with mHSPC between 2018-2024. PSA response at 6-12 months and volume of disease were determined. DNA alterations were classified by Somatic Tumor Risk Assessment for Overall Survival-Prostate (STRATOS-P) genomic system, a validated method of risk stratification. PSA response was grouped into ≤0.2, &gt;0.2-&lt;2, 2-&lt;10, and ≥10 ng/mL. Charlson comorbidity index (CCI) high was defined as ≥3. PSMA-PET based staging was determined if PSMA-PET performed within 100 days of mHSPC diagnosis. Multivariable Cox model-based weights were used to create the STRATOS-P mHSPC clinical risk score to prognosticate overall survival and select patients for intermittent therapy. Kaplan-Meier analysis was used to estimate Overall Survival (OS) and time to death or castration resistance from diagnosis, a real-world progression-free survival (rwPFS) surrogate. Results: In 3094 Veterans identified, 937 (30.3%) had DNA-based genomic testing within 6 months of diagnosis. There were 1349 patients (43.6%) with a PSA of ≤0.2 who had a median OS of 74.6 months. Of patients with genomic testing, 459 (49.0%) were STRATOS-P Clinical low risk with a median OS that was not reached and rwPFS of 72.1 months, moderate risk with median OS of 37.1 months and rwPFS of 21.6 months, and high risk with median OS of 19.1 months and rwPFS of 12.1 months. There were 66 (13.0%) patients that were STRATOS-P Clinical low risk but did not achieve ≤0.2% and 116 (22.8%) patients that achieved PSA ≤0.2 but were considered STRATOS-P clinical moderate risk. Approximately 497 patient (53%) could be considered for intermittent therapy and had a median OS of 77.9 months and rwPFS of 56.2 months, similar to patients with PSA≤0.2. Conclusions: STRATOS-P Clinical model in mHSPC is prognostic for OS and rwPFS and can identify patients for intermittent therapy to achieve similar OS and rwPFS compared to selection using PSA ≤0.2 at 6-12 months. Future trials should incorporate clinico-genomic assessments to improve risk stratification and selection of therapy. STRATOS-P clinical model (n=937). n HR 95.0% CI CCI High 397 1.62 1.34-1.95 High Volume 608 1.53 1.31-1.78 STRATOS-P Genomic risk* Intermediate 470 1.53 1.23-1.92 Unfavorable 132 2.39 1.80-3.17 Non-PSMA-PET staging 708 1.62 1.23-2.14 &gt;0.2 - 2 210 1.60 1.25-2.04 PSA Response# &gt;2 - &lt;10 131 2.98 2.30-3.84 10+ 87 5.64 4.26-7.47 *STRATOS-P genomic risk favorable is referent, n=335. #PSA Response ≤0.2 is referent, n=509.