Browse Articles

Immune-related adverse events and associated clinical outcomes in cancer patients treated with immune checkpoint inhibitors: A multicenter real-world study.

e24149 Background: Immune checkpoint inhibitors (ICIs) are widely used in cancer treatment to boost the immune system's ability to target cancer cells. However, this can also induce immune-related adverse events (irAEs), which may impair organ function and affect patients' quality of life. Understanding the real-world impact of these irAEs is crucial. The primary outcome was to estimate the incidence and timing of side effects of ICIs. The secondary outcome was to assess and identify the pattern of irAEs and patient-disease-related factors associated with irAEs, time to onset of irAEs, and clinical outcomes. Methods: We conducted multi-centre retrospective cohort study in two large tertiary care centres at Riyadh, Saudi Arabia. We examined the use of ICIs in cancer patients aged 18 and older who received at least one dose of ICI between January 2017 and January 2023. Patients were followed until July 2025, death, or last documented clinical follow-up, whichever occurred first. Data, including demographics, clinical details, and treatment information, were extracted from electronic medical records (EMR). Results: A total of 608 patients were included (median age 62 [50–71] years; 55.9% male). The most common cancers were lung (118; 19.4%) and liver (85; 14.0%). Stage IV disease was present in 469 (77.1%), and metastases in 454 (74.7%). Pembrolizumab (307; 50.5%) and Nivolumab (185; 30.4%) were the most frequently used ICIs; and 305 (50.2%) patients received combination chemotherapy. Overall, 360 (59.2%) patients developed at least one irAE. The most frequent irAEs were fatigue (145; 23.8%), endocrine irAEs (93; 15.3%, mainly hypothyroidism 87; 14.3%), skin reactions (62; 10.2%), and colitis (54; 8.9%). Median time to irAEs onset was 84 [37–181] days; neurologic irAEs occurred earliest 20 [9–74] days, and endocrine irAEs occurred latest 135 [82–252] days. Disease progression occurred in 351 (57.7%) and was more frequent without irAEs (63.7% vs 53.6%, p =0.0132). The median progression-free survival (PFS) was 13.53 months (95%CI, 11.04–19.12) with irAEs versus 5.95 months (95%CI, 4.37–7.92) without irAEs (log-rank p <0.0001). The median overall survival (OS) was 23.85 months (95%CI, 18.73–37.88) with irAEs versus 17.67 months (95%CI, 14.06–28.94) without irAEs (log-rank p =0.176). Conclusions: In this large real-world study, irAEs were prevalent among patients receiving ICIs and showed variation in both time and organ involvement. Patients who experienced irAEs demonstrated prolonged PFS, although OS was numerically prolonged but did not achieve statistical significance. Overall, these findings highlight the clinical importance of recognizing and monitoring irAEs during ICI treatment in routine practice and its clinical implications.

Hydrothermal synthesis and multifunctional characterization of cobalt ferrite–rGO nanocomposites for theranostic applications

Methanol‐Ethanol Discrimination and Selective Sensing Enabled by Molecular Sieving in Conductive MOFs

ABSTRACT Methanol presents a significant health risk because of its volatility and toxicity. Its close chemical similarity to ethanol increases the likelihood of accidental ingestion through contaminated beverages. Here, we report chemiresistive sensors that selectively distinguish methanol from ethanol under ambient conditions. The sensors consist of single‐walled carbon nanotubes (CNTs) functionalized with conductive metal–organic frameworks (cMOFs) constructed from 2,3,7,8,12,13‐hexahydroxytetraazanaphthotetraphene (HHTT). Trinuclear intra‐pore clusters (IPCs) located within the honeycomb channels of HHTT‐based cMOFs govern methanol sensing by increasing the density of adsorption sites and constraining molecular diffusion through the pores. Mg‐HHTT, in which the pores are largely occupied by IPCs, enhances the methanol response of CNT@cMOF composites while suppressing transport of ethanol and larger alcohols. In contrast, isostructural Ni‐HHTT and Cu‐HHTT analogues, which lack a high density of IPCs, exhibit substantially lower sensitivity and selectivity. Density functional theory and molecular dynamics simulations support a sensing mechanism based on IPC‐mediated molecular sieving. Sensor tests using methanol‐spiked liquors demonstrate selective methanol detection in complex beverage matrices under practical ambient conditions.

Association of concurrent GLP-1 receptor agonist use with survival outcomes in patients with metastatic colorectal cancer receiving immune checkpoint inhibitors.

3605 Background: Metabolic status and systemic inflammation are increasingly recognized as important modifiers of immune checkpoint inhibitor (ICI) response. Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) exhibit anti-inflammatory and immunomodulatory properties and are widely used in diabetes and obesity. However, their clinical impact on immunotherapy outcomes in metastatic colorectal cancer (mCRC) remains unclear. Methods: We conducted a retrospective observational cohort study using the TriNetX Global Collaborative Network, including electronic health records from 170 healthcare organizations between 2015 and 2025. Adults with mCRC who received at least one ICI were identified. Patients prescribed GLP-1 RAs within six months before or after ICI initiation were compared with non-exposed controls. Patients with prior bariatric surgery were excluded. Cohorts were balanced using 1:1 propensity score matching for demographics, metabolic conditions, and major comorbidities. Primary endpoint was all-cause mortality at 1, 3, and 5 years. Secondary outcomes included pneumonia, heart failure exacerbation, and major abdominal surgical procedures. Survival analyses were performed using Kaplan–Meier methods and Cox proportional hazards models. Results: Among 8,304 eligible patients, 148 received GLP-1 RAs. After propensity matching, 138 patients were included in each cohort with well-balanced baseline characteristics. At 1 year, mortality was lower in the GLP-1 cohort (27.5% vs 40.6%), corresponding to a 32% relative risk reduction (RR 0.68, 95% CI 0.49–0.95) and reduced hazard of death (HR 0.67, 95% CI 0.43–1.03). This association persisted at 3 years (27.5% vs 41.3%; RR 0.67, 95% CI 0.48–0.93; HR 0.68, 95% CI 0.45–1.05) and 5 years (28.3% vs 42.0%; RR 0.67, 95% CI 0.49–0.93), with higher 5-year survival probability (50.5% vs 41.5%). Secondary outcomes favored GLP-1 exposure, including lower incidence of pneumonia (10.9% vs 17.4%; HR 0.58) and fewer heart failure exacerbations (23.9% vs 31.9%; HR 0.75). Rates of major abdominal surgical procedures were numerically lower in the GLP-1 cohort. Conclusions: In this large real-world cohort, concurrent GLP-1 RA exposure was associated with clinically meaningful and durable survival benefit in mCRC patients receiving ICIs, without increased adverse clinical outcomes. These findings suggest a potential immunometabolic interaction and support prospective evaluation of GLP-1–based strategies as adjunctive modifiers of immunotherapy response.

Impact of treatment sequence of immunotherapy and stereotactic radiotherapy on survival in non–small cell lung cancer patients with brain and bone metastases: An NCDB data analysis.

e20119 Background: Optimizing the sequence of immunotherapy (IT) and stereotactic radiotherapy (SRT) in metastatic non-small cell lung cancer (NSCLC) may enhance survival, yet site-specific efficacy remains undefined. We investigated the impact of IT followed by SRT (IT→SRT) versus SRT followed by IT (SRT→IT) sequencing on overall survival (OS) in NSCLC patients with brain metastases (BrMs) or bone metastases using National Cancer Database (NCDB) data. Methods: This retrospective cohort study included 3,158 adults with Stage IV NSCLC diagnosed between 2015 and 2022 who had either brain metastases (n = 2,475) or bone metastases (n = 683). All patients received both SRT and IT, started 10–66 days apart. Patients were grouped by treatment sequence: SRT→IT or IT→SRT. Confounding was addressed using inverse-probability-of-treatment weighting (IPTW) and overlap weighting (OW) derived from propensity scores. OS was evaluated using Kaplan–Meier estimation and Cox proportional-hazards models within each metastasis site. Missing covariates were handled with multiple imputation by chained equations (25 datasets), and hazard ratios (HRs), 95% confidence intervals (CIs), and p values were combined across imputations using Rubin’s rules. Results: In the brain-metastasis cohort, 2,061 patients received SRT→IT and 414 received IT→SRT. Median OS was 27.4 months with SRT→IT versus 23.3 months with IT→SRT. The MI-pooled unweighted Cox model yielded an HR of 1.13 (95% CI 0.98–1.30; p = 0.081) for IT→SRT versus SRT→IT, while IPTW weighting produced a similar but statistically significant association (HR 1.10, 95% CI 1.02–1.19; p = 0.012); OW estimates were directionally similar but less precise (HR 1.10, 95% CI 0.90–1.35; p = 0.341). In the bone-metastasis cohort, 520 patients received SRT→IT and 163 received IT→SRT. IT→SRT was associated with longer survival (median OS 22.7 vs 15.5 months). The MI-pooled unweighted Cox HR for IT→SRT versus SRT→IT was 0.81 (95% CI 0.64–1.01; p = 0.064), with a statistically significant effect under IPTW weighting (HR 0.79, 95% CI 0.68–0.91; p = 0.001) and a directionally consistent but less precise OW estimate (HR 0.80, 95% CI 0.58–1.10; p = 0.165). Conclusions: Treatment-sequencing effects in metastatic NSCLC appear to be site specific. In patients with brain metastases, SRT→IT was associated with modestly longer OS, with sequence effects sensitive to the weighting approach. In patients with bone metastases, IT→SRT was associated with substantially longer OS, corresponding to an approximate 20% relative reduction in mortality in IPTW-weighted analyses. These findings support consideration of metastasis-site biology when determining the sequencing of IT and SRT and highlight the need for prospective trials explicitly designed to evaluate site-tailored multimodality strategies.

Nanoscale spatial profiling of DLL3, SEZ6, and B7-H3 in small cell lung cancer using a rapid expansion microscopy assay.

e15194 Background: Small cell lung cancer (SCLC) remains an aggressive malignancy with limited therapeutic options. Recent advances have expanded the treatment landscape beyond chemotherapy and immune checkpoint inhibitors, most notably with the DLL3-directed bispecific T-cell engager tarlatamab, which has demonstrated clinically meaningful activity in relapsed disease and is now moving into earlier-line trials. In parallel, antibody–drug conjugates (ADCs) targeting SEZ6 and B7-H3 (CD276) have shown promising early clinical signals. For these emerging targets, both expression level and subcellular localization may influence therapeutic efficacy, particularly for ADC payload delivery and T-cell engagement. However, SCLC diagnostic specimens are frequently limited, and conventional immunohistochemistry lacks the multiplexing capacity and spatial resolution needed to extract this information. We evaluated a rapid, clinically optimized expansion microscopy (ExM) approach to enable multiplexed nanoscale profiling from scarce SCLC tissue. Methods: We applied a streamlined Magnify ExM protocol with a total turnaround time of approximately 6 hours, designed for compatibility with routine pathology workflows. Key steps were partially automated using an Opentrons Flex platform to improve reproducibility. Formalin-fixed paraffin-embedded (FFPE) SCLC tissues were physically expanded to enable super-resolution imaging on standard diffraction-limited microscopes. To allow simultaneous tri-marker analysis on limited tissue, we used Proteintech FLEXABLE 2.0 antibody labeling kits, enabling concurrent staining with three rabbit primary antibodies (CD276, SEZ6, DLL3) on the same tissue section without species cross-reactivity. Tissue microarrays from 24 SCLC patients (ages 28–75; stages I–IIIA) were analyzed. Results: Magnify ExM preserved tissue architecture and protein integrity while achieving effective spatial resolution of ~60–70 nm. Multiplexed imaging revealed marked inter- and intra-patient heterogeneity that was not detectable by conventional methods. Distinct nanoscale clustering patterns of B7-H3 and SEZ6 were observed, with potential implications for ADC internalization, while DLL3 demonstrated variable surface-to-Golgi localization ratios that may differentiate suitability and response for ADCs versus T-cell engagers. FLEXABLE 2.0 enabled dense multiplexing on rare biopsies, substantially increasing data yield per specimen. Conclusions: This rapid Magnify-based assay enables clinically feasible, nanoscale spatial profiling of emerging SCLC targets from limited tissue. By resolving subcellular organization of CD276, SEZ6, and DLL3, this approach provides a path toward potentially more informed patient stratification for ADC- and bispecific-based therapies while maximizing diagnostic value from scarce clinical specimens.

The global and regional burden of leukemia from 1990-2023: A systematic analysis for the Global Burden of Disease study 2023.

e18625 Background: Leukemia represents an important component of global hematologic malignancy burden across the age spectrum, particularly for children and adolescents. While advancements in diagnostics and treatment have the potential to reduce the global burden from leukemia, it remains unclear how effective these gains have been distributed across all regions. This study aims to analyze leukemia incidence and mortality from 1990-2023 globally as well as by World Bank income group, age group, and leukemia type. Methods: Data from population-based cancer registries and vital registration systems were used to generate ensemble models for mortality counts and age-standardized rates for leukemia and five leukemia types: acute lymphoid leukemia (ALL), acute myeloid leukemia (AML), chronic lymphoid leukemia (CLL), chronic myeloid leukemia (CML), and other leukemias. Matched incidence and mortality data were used to generate modeled mortality to incidence ratios, which were subsequently applied to mortality estimates to generate incidence. Results: Leukemia was the ninth most common cancer globally in 2023 among all ages with 582,000 (450,000 to 745,000) incident cases and 342,000 (307,000 to 382,000) deaths. Among 0-19 year-olds, there were 98,600 (68,000-139,000) incident cases and 45,900 (38,700 to 56,200) deaths globally, making it the most common cancer in children. Globally, AML was the most common leukemia subtype among all ages in 2023 while ALL was greatest among children aged 0-19 years. Between 1990 and 2023, global leukemia all-age age-standardized mortality rates (ASMR) decreased by 33.2% (24.2 to 41.0) from 5.9 (5.3 to 6.5) to 4.0 (3.6 to 4.4) per 100,000 while all-age age-standardized incidence rates (ASIR) changed by -21.4% (-41.8 to 10.7) from 8.7 (7.3 to 10.6) to 6.8 (5.2 to 8.8) per 100,000. The global 2023 0-19 ASIR was 3.8 (2.6 to 5.4) per 100,000 and the ASMR was 1.7 (1.5 to 2.1). Reductions in ASMRs between 1990 and 2023 were most pronounced in upper middle income and high income countries, -40.3 (-47.8 to –28.1) and –29.3 (-33.5 to –24.8), respectively, compared to the changes seen in lower middle income and low income countries, -18.8 (-41.8 to 10.7) and –19.5 (-43.6 to 16.0), respectively. Conclusions: Estimates from GBD 2023 demonstrate that leukemia remains the most common childhood cancer with leukemia type patterns varying across adults and children. Although ASMRs are declining globally, there have been more substantial declines in higher income countries compared to lower income countries, calling attention to inequities. These age and geographic patterns underscore the importance of targeted approaches to reducing the burden of leukemia and increased efforts to expand access to effective diagnostic and treatment services in lower resourced settings.

Radiotherapy in symptomatic metastatic triple-negative breast cancer treated with first-line pembrolizumab and chemotherapy: A multinational real-world study.

e13112 Background: Despite the addition of pembrolizumab to chemotherapy, the prognosis of patients with symptomatic metastatic triple-negative breast cancer (mTNBC) remains poor. We evaluated the indications, toxicity, and outcomes of radiotherapy (RT) added to pembrolizumab and chemotherapy in real-world mTNBC. Methods: This multinational, multicenter retrospective study was conducted within the CEBCC-101 real-world evidence project across 18 cancer centers. Female patients with histologically confirmed mTNBC treated with first-line pembrolizumab plus chemotherapy outside clinical trials in Poland, the Czech Republic, and Slovakia between September 2022 and June 2025 were included. Among 178 eligible patients, 54 received RT. Stereotactic radiotherapy was used in 20 cases. Results: RT was primarily administered for symptomatic disease, most commonly pain, including cases with actual or impending pathological fractures or risk of spinal cord compression. Additional indications included irradiation of progressing sites, superior vena cava syndrome, and bleeding from ulcerated breast tumors. Only two patients received RT for oligometastatic disease. RT sites included bone (n = 20), central nervous system (n = 19), breast/chest wall or regional lymph nodes (n = 10), and other sites (n = 5). Median local control was 14.6 months. RT was generally well tolerated, with acute toxicities reported in 9 patients (16.7%), predominantly grade 1–2, and one grade 3 event. Patients receiving RT had higher rates of hematologic adverse events (67.9% vs 32.1%; p = 0.045) and alopecia (69.8% vs 30.2%; p = 0.029), with no significant differences in chemotherapy dose reductions, treatment delays, or other systemic toxicities. Given the high-risk clinical characteristics of patients requiring RT, progression-free survival (PFS) and overall survival (OS) were significantly shorter in this group. Median PFS was 7.4 months in patients receiving RT versus 8.6 months in those without RT ( p = 0.007), with 12-month PFS rates of 20.3% and 43.5%, respectively. Median OS was 15.6 versus 26.8 months ( p = 0.002), with 24-month OS rates of 22.7% and 56.0%, respectively. Conclusions: Patients with symptomatic mTNBC requiring radiotherapy represent a high-risk population with poor outcomes despite combined pembrolizumab and chemotherapy. While radiotherapy provides effective local control with acceptable toxicity, prospective studies are needed to determine whether greater integration of stereotactic radiotherapy with antibody–drug conjugates can improve outcomes in this setting.

Enhancing OP-35 classification of potentially avoidable hospital visits after chemotherapy using procedure codes.

e23075 Background: Medicare’s OP-35 quality measure defines potentially avoidable hospital visits within 30 days of outpatient chemotherapy using the first non-malignancy discharge diagnosis code. Prior work has found poor diagnostic characteristics of OP-35 compared to clinician-reviewed hospital visit (sensitivity 35%, specificity 71%, accuracy 60%, AUROC of 0.53). A single discharge diagnosis code may inadequately capture the clinical complexity of a hospital visit. Using the random sample of clinician-reviewed hospital visits from the prior study, we developed a claims-based definition for potentially avoidable hospital visits incorporating procedure codes as well. Methods: We analyzed 705 acute hospital visits (5% random sample of 12,597 hospital visits) occurring within 30 days of chemotherapy that underwent blinded clinician chart review, with visit avoidability adjudicated by majority agreement and used as the definition for true avoidability. We constructed encounter-anchored timelines linking all day-level CPT, HCPCS, and ICD-9 procedure codes across the hospital stay (50,169 procedure codes from the 705 visits), and derived interpretable features reflecting care intensity and timing, including procedure diversity, temporal clustering and repeated procedures, and encounter length of stay. We applied multiple procedure-augmented classification approaches, including linear, nonlinear, and pattern-based models, to predict clinician-adjudicated visit avoidability. Model development used five-fold cross-validation for tuning and internal validation, followed by evaluation on a held-out test set comprising 30% of encounters. Performance was evaluated using sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUROC). Results: Of the 705 hospital visits, clinicians classified 213 visits (30.2%) as potentially avoidable. Our procedure-code augmented OP-35+ method demonstrated substantially improved discrimination, with test-set AUROC ranging from 0.86 to 0.88. Across evaluated approaches, sensitivity improved to approximately 64–66% while maintaining high specificity (88–90%), yielding overall accuracy near 79%. Features capturing procedure intensity, multi-day service patterns, and interactions with length of stay consistently contributed most to model performance and improved differentiation between outpatient-manageable encounters and visits requiring urgent or emergent inpatient care. Conclusions: Incorporating procedure codes and encounter-level service patterns improved alignment with clinician-defined avoidable hospital visits after chemotherapy. A procedure-informed OP-35+ definition provides a more clinically grounded approach to identifying avoidable hospital visits in claims data, with implications for cancer care quality measurement and policy.

Predictive biomarkers for PFS in patients receiving quaratusugene ozeplasmid.

e15184 Background: Quaratusugene ozeplasmid is a gene therapy that delivers a plasmid coding for the TUSC2 tumor suppressor gene to lung cancer cells, as > 80% of lung cancers have been shown to have decreased or absent TUSC2 protein. TUSC2 protein levels have not correlated with PFS, presumably because of the complexities of TUSC2 protein regulation. Preclinical studies have identified higher levels of Trop-2 protein in organoids and lower levels of PTEN protein in lung cancer cell lines as correlating with response (AACR 2026). Tumor tissue from patients in clinical trials with quaratusugene ozeplasmid were evaluated for Trop-2 and PTEN protein expression. Methods: Monoclonal antibodies against Trop-2 (BSB148 from BioSB) and PTEN (138G6 from Cell Signaling Technology) were used for immunohistochemistry in paraffin sections from archival tumor samples in patients enrolled in 3 clinical trials with quaratusugene ozeplasmid and results expressed as H-scores. H-scores were calculated by evaluating diaminobenzidine staining intensity using the formula [1 × (% cells 1+) + 2 × (% cells 2+) + 3 × (% cells 3+)]. Results: Data on Trop-2 and PTEN protein expression and data on Progression Free Survival (PFS) were available from 18 patients enrolled in clinical trials with quaratusugene ozeplasmid. Six patients with NSCLC were enrolled in the Acclaim-1 trial in combination with osimertinib, and one was enrolled in the Acclaim-2 trial in combination with pembrolizumab. Eleven patients with SCLC were enrolled in the Acclaim-3 trial in combination with atezolizumab. In patients with NSCLC, Trop-2 H-scores above 100 correlated with prolonged PFS (p = 0.05), and PTEN H-scores below 100 correlated with prolonged PFS (p = 0.03). In patients with SCLC, Trop-2 H-scores were universally low, and thus non-evaluable. PTEN H-scores in patients with SCLC did not correlate with prolonged PFS (p = 0.53). Conclusions: Following up on preclinical cell line and organoid models indicating that Trop-2 and PTEN protein expression correlated with response, levels of Trop-2 and PTEN protein were evaluated in patients treated with quaratusugene ozeplasmid. Despite the small number of samples evaluated, both Trop-2 H-scores above 100 and PTEN H-scores below 100 correlated with longer PFS in patients with NSCLC, but not in patients with SCLC. Clinical trial information: NCT04486833 ; NCT05062980 ; NCT05703971 .

ALPACCA: Phase 3 trial of firmonertinib vs investigator’s choice of EGFR inhibitor as first-line treatment for locally advanced or metastatic NSCLC with EGFR P-loop and alpha c-helix compressing (PACC) uncommon mutations (FURMO-006).

TPS8675 Background: Despite advances in treatment of NSCLC with EGFR classical mutations, no universally accepted standard-of-care treatment exists for patients with EGFR uncommon mutations including PACC mutations. Treatment options include EGFR tyrosine kinase inhibitors (TKIs) (osimertinib, afatinib), chemotherapy, or other targeted therapies (amivantamab for exon 20 insertion mutations). EGFR PACC mutations comprise approximately 12.5% of all NSCLC EGFR mutations (Robichaux et al 2021, Nilsson et al 2024). Firmonertinib is a once daily oral, highly brain penetrant, broadly active mutant-selective EGFR inhibitor that targets classical and uncommon mutations (Musib et al., NACLC 2022). In the phase 1b FURMO-002 study (FURTHER), first-line locally advanced or metastatic EGFR PACC mutation NSCLC patients treated with firmonertinib 240 mg daily achieved a confirmed ORR of 68.2%, best ORR of 81.8%, disease control rate (DCR) of 100%, and median progression-free survival (mPFS) of 16.5 months by blinded independent central review (BICR) (Le et al., WCLC 2025). Firmonertinib was generally well-tolerated with manageable EGFR TKI-associated adverse events. Methods: ALPACCA (FURMO-006; NCT07185997) is a global, phase 3, randomized, open-label study investigating firmonertinib vs investigator’s choice of osimertinib or afatinib. Eligible patients have locally advanced or metastatic NSCLC with EGFR PACC mutations. Key inclusion criteria include documented presence of EGFR PACC mutation and measurable disease per RECIST v1.1. Patients with asymptomatic CNS metastases are allowed. Key exclusion criteria include prior systemic anticancer therapy in the locally advanced or metastatic setting or any prior EGFR TKI therapy. Approximately 480 patients will be randomized 1:1 to receive firmonertinib 240 mg daily or investigator’s choice of osimertinib 80 mg daily or afatinib 40 mg daily. Primary endpoints are PFS and ORR per RECIST v1.1 by BICR. Key secondary endpoints include OS, investigator assessed PFS and ORR, and safety and tolerability. Enrollment is ongoing. Clinical trial information: NCT07185997 .

Survival differences in advanced gynecological cancers after the development of immune checkpoint inhibitors and targeted therapy: United States SEER-based population study.

e17652 Background: Immune checkpoint inhibitors (ICI) are being used for advanced gynecological cancers (cervix, endometrium, and ovary). Antiangiogenic (AA) agents/chemotherapy (Chem) and PARP inhibitors (PARPi) were approved in 2014. ICI was approved in 2017. ICI±AA/PARPi/Chem (comb) has been used since 2017. We aim to study survival differences before and after these approvals. Methods: We used Surveillance, Epidemiology and End Results (SEER) research plus database (Nov 2024 submission,17 registries). Patients aged ≥ 20 years; all histology of cervix, uterus and ovarian cancer with “distant” (combined summary stage) from 2011 to 2019 were included. Time was stratified as 2011-2013, 2014-2016, 2017-2019. OS in these time periods was calculated using Kaplan Meier (KM) method and was compared with log rank tests. Multivariate analysis was done using cox proportional hazard ratios, and p < 0.05 was considered for statistical significance. IBM SPSS was used for analysis. Results: Out of 36535 patients, 4172 had cervical cancer, 9832 had endometrial cancer, and 22531 had ovarian cancer. All years of diagnosis had 5-year survival except for 2019, which had 4-year survival. The survival difference among three timelines was statistically significant (log rank test, χ2 = 8.128, p=0.017). Stratified analysis showed cervical cancer (χ2 = 7.781, p = 0.020) and endometrial cancer (χ2 = 11.36, p = 0.003) had improved survival except ovarian cancer (χ2 = 0.615, p=0.735), across these timelines. On multivariate analysis, 2011-2013 had a significantly higher risk of dying compared to 2017-2019. 2014-2016 had no significant difference in risk of dying compared to 2017-2019 (see table). Conclusions: In this population-based study, use of ICI showed survival benefits in advanced gynecological cancers of cervical and uterine origin, supporting clinical trials. Increased mortality risk in non-Hispanic Black race, older age ≥40 years, divorced/widowed status should guide further measures. Elevated CA125 increases mortality risk. Survival in ovarian cancer is not promising, but future analysis with new databases that include recent trials will help show changing survival. Multivariate analysis. Variables HR (95% CI) p-value Age <40 y 0.629 (0.591-0.669) <0.001 ≥40 y Ref Year group 2011-2013 1.043 (1.012-1.074) 0.006 2014-2016 1.028 (0.998-1.059) 0.064 2017-2019 Ref Race/origin Non-Hispanic White Ref Non-Hispanic Black 1.271 (1.228-1.316) <0.001 Non-Hispanic Asian/Pacific Islander 0.839 (0.803-0.876) <0.001 Hispanic 0.898 (0.869-0.929) <0.001 Chemotherapy Yes 0.38 (0.37-0.39) <0.001 No/unknown Ref CA125 status (Ovary) Positive/Elevated 1.182 (1.028-1.360) 0.019 Negative Ref Marital status Single/unmarried 0.923 (0.868-0.981) 0.01 Married 0.794 (0.749-0.842) <0.001 Divorced/widowed 1.128 (1.063-1.198) <0.001 Unknown Ref

Phase Ib expansion and dose optimization study of CX-5461 in patients with solid tumours enriched for DNA-repair deficiencies.

3091 Background: G-quadruplexes (G4) are guanine-rich DNA/RNA structures that contribute to DNA damage and genomic instability. Pidnarulex (CX-5461) is a first-in-class G4 stabilizer that induces synthetic lethality in Homologous Recombination-deficient (HRD) cells. Previous phase 1 study showed safety and preliminary activity. Here we report dose optimization for recommended phase 2 dose (RP2D) of CX-5461 and confirm safety. Methods: CX-5461 was administered intravenously on Days 1 and 8 of a 28-day cycle at 2 dose levels (250 and 325 mg/m²) in patients (pts) with advanced pancreatic adenocarcinoma (PA), breast cancer (BC), or ovarian cancer (OC). The main cohort (Arms A: 250 mg/m² & B: 325 mg/m²) included pts with HRD or DNA damage response (DDR) alterations and has completed accrual. An exploratory cohort (Arms C: 250 mg/m² & D: 325 mg/m²) of OC pts with BRCA or other HRD genes enrolled simultaneously. Primary objective is to determine RP2D of CX-5461 using Relative Dose Intensity (RDI). Secondary objectives are safety, tolerability and efficacy. Results: 54 pts were treated and evaluable for toxicity (Arm A = 18, Arm B = 18, Arm C = 10, Arm D = 8). Patients’ characteristics and efficacy results are detailed in table 1. The median number of prior lines of treatment were 6 (1-14). The median number of CX-5461 cycles received were 2 (1-30) and weeks on treatment were 8 (4-120). Grade ≥3 Treatment-Related Adverse Events (TRAEs) occurred in 12 pts (22%), including proteinuria, thrombocytopenia, and fatigue; 6 (11%) were AEs of special interest: 1 (2%) palmar-plantar erythrodysesthesia, 4 (7%) photosensitivity, and 1 (2%) phototoxic drug eruption. Among 29 response-evaluable pts in main cohort, 1 had PR (3%) and 10 had SD (34%),median duration 16 weeks (wks) (8-84). In evaluable OC (n=29), 2 had PR (7%), 13 SD (44%), median duration 16 wks (8-84). OC pts with BRCA mutations (n=25) had 2 PR (8%) and 11 SD (44%); all had prior PARPi exposure. 3/ 5 non-BRCA HRD had SD (60%). Prolonged clinical benefit was observed in 4 pts with treatment durations of 14 (OC, gBRCA1), 17 (OC, gBRCA2), 21 (OC, sBRCA1) and 30 cycles (PA, gPALB2). RDI was 91% in Arm A and 88% in Arm B. RDI and clinical data support 250 mg/m² as the RP2D. Conclusions: CX-5461 was tolerable and showed evidence of durable disease control in heavily pretreated pts, enriched for HRD/DDR alterations and post exposure to PARPi. The 250 mg/m² dose is established as the RP2D based on clinical data and RDI. Correlative studies are ongoing. Clinical trial information: NCT04890613 . Arm A (N=18) Arm B (N=18) Arm C (N=10) Arm D (N=8) Tumor Type BC 7 (39%) 2 (11%) - - OC 5 (28%) 9 (50%) 10 (100%) 8 (100%) PC 6 (33%) 7 (39%) - - Molecular alterations gBRCA1/2 12 (67%) 9 (50%) 3 (30%) 3 (38%) sBRCA1 0 (0%) 1 (6%) 6 (60%) 3 (38%) Other 6 (34%) 8 (44%) 1 (10%) 2 (25%) Efficacy (evaluable pts only) PR 1/13 (8%) 0 (0%) 0 (0%) 1/7 (14%) SD 4/13 (31%) 6/16 (38%) 5/9 (56%) 3/7 (43%) Disease control rate 5/13 (39%) 6/16 (38%) 5/9 (56%) 4/7 (57%)

TROP2 as an actionable biomarker for anal cancer.

3519 Background: Chemoimmunotherapy is standard frontline therapy for metastatic anal cancer, with no effective treatment options afterward. Mutation profiling has not yielded matched targeted therapies against anal cancer. Trophoblast cell-surface antigen 2 (TROP2) is overexpressed in other solid tumors, and TROP2 antibody-drug conjugates (ADCs) have demonstrated efficacy in breast and lung cancers. Given the therapeutic availability against TROP2, we evaluated expression of TROP2 in anal cancer as an actionable therapeutic target. Methods: Metastatic anal cancer tumors from 33 patients at MD Anderson were sequenced by bulk RNA sequencing (“Tumor Portrait” assay, Boston Gene) to quantify TACSTD2 (TROP2) gene expression. TACSTD2 expression was compared across all solid tumors and classified according to ranked expression percentile: high (>83 rd percentile), medium (17-83 rd percentile), or low (<17 th percentile). Gene expression (Nanostring) for >18,000 genes was measured using whole-genome digital spatial profiling (DSP) on a separate cohort of 40 chemoradiotherapy-refractory localized anal cancers collected at salvage surgery. TACSTD2 gene expression on tumor cells vs TME cells was compared with a t-test (SPSS). TROP2 protein expression (Invitrogen) was quantified by H-score using immunohistochemistry (IHC) and correlated with TACSTD2 expression via Spearman’s correlation. Three patient-derived xenograft (PDX) models of anal cancer (C1411, C1436, O0026) were treated with normal saline (untreated control, UTC) or the TROP2 ADC sacituzumab govitecan (sac-gov; 10 mg/kg IP twice weekly). Tumor growth inhibition (TGI) was defined as 1 - (mean tumor volumes of sac-gov/UTC) at 21 days. Results: Median TACSTD2 (TROP2) gene expression for patients with metastatic anal cancer was ranked at the 75 th percentile (IQR 63-85) relative to all solid tumors: 13/33 (39%) with high, 20/33 (61%) with medium, and none with low TACSTD2 expression. For localized anal cancers, TACSTD2 gene expression using DSP was significantly higher on tumor segments vs TME segments (log fold change 3.49, adjusted p < 0.0001). 30/39 (77%) anal cancers had high TROP2 expression by IHC, defined by H-score >200. Correlation between TACSTD2 gene expression and TROP2 protein expression by IHC was observed (r= 0.43, p < 0.001). In PDX models, TGI with sac-gov relative to UTC was observed in both models with high TROP2 expression (C1411: TGI 54%, p< 0.001; C1436: TGI 52%, p< 0.001) but not in a model with no TROP2 expression (O0026: TGI 12%, p=n.s.). Conclusions: TROP2 expression is high for localized and metastatic anal cancer. TACSTD2 gene expression correlates with matched TROP2 protein expression for anal cancer. Anti-tumor efficacy of TROP2 ADCs in vivo supports a forthcoming trial of sacituzumab tirumotecan for patients with treatment-refractory metastatic anal cancer, with the ultimate goal of establishing TROP2 as a predictive biomarker for treatment benefit in this rare cancer.

Association between the rs72725854(A>T) allele and prostate cancer in males of African ancestry in a US population-based database.

10504 Background: Black males have a 70% excess incidence of and two-fold higher mortality from prostate cancer (PCa) compared to White males in the US. The rs72725854(A>T) variant (‘T’ allele) is a common African ancestry-specific variant that is associated with PCa in large case-control studies in the US and Africa. We measured the association between the ‘T’ allele and PCa in a unique US population-based cohort. Methods: We conducted a retrospective time-to-event cohort study using the All of Us database (Duke IRB Pro00117610). We included males of African ancestry who met the following criteria: 1) were at least 35 years old at All of Us enrollment; 2) had undergone whole genome sequencing; 3) had linked electronic health record (EHR) data; 4) had no Systemized NOmenclature of MEDicine – Clinical Terms (SNOMED-CT) diagnosis of PCa prior to All of Us enrollment. PCa diagnoses were ascertained by the following EHR-based phenotyping algorithm: two PCa SNOMED-CT diagnosis codes plus evidence of PCa treatment with radical prostatectomy, radiotherapy or androgen deprivation therapy. The date of PCa diagnosis was defined as the latter of second diagnosis code date or first treatment date. Individuals were censored at last visit in the EHR if they did not meet the primary endpoint of PCa diagnosis. The association between each additional ‘T’ allele and time from All of Us enrollment to PCa diagnosis was measured using Cox regression methods, adjusting for the top 5 genetic principal components, age, and rare pathogenic variants (RPVs) in PCa predisposition genes ( ATM, BRCA1/2, HOXB13 X285K African-ancestry variant , and PALB2 ). Included RPVs had pathogenic/likely pathogenic designation in ClinVar with at least two stars. Results: Our study included 18,846 male participants of African ancestry, with 2,349 (12.5%) and 90 (0.5%) being heterozygous and homozygous for the ‘T’ allele, respectively. Median age at enrollment was 56.0 years (interquartile range 48.5 – 62.3). There were 91 cases of PCa during a median follow-up of 25.6 months (interquartile range 3.0 – 45.9). For each additional ‘T’ allele, there was a 3.4-fold increased hazard of PCa (HR 3.38; 95% CI 2.31 – 4.95; p 3.92x10 -10 ). Age (HR 1.09; 95% CI 1.06 – 1.11; p 2.74x10 -15 ) and RPVs in ATM (HR 7.54; 95% CI 1.84 – 30.90; p 5.03x10 -3 ) were also significantly associated with increased PCa diagnosis. BRCA2 RPVs (HR 4.94; 95% CI 0.68 – 35.68; p 0.112) and HOXB13 X285K (HR 2.61; 95% CI 0.36 – 19.02; p 0.343) were associated with PCa, but did not meet statistical significance. There were no BRCA1 or PALB2 RPVs among participants with PCa. Conclusions: In a large, US population-based study, we confirmed an association between the common rs72725854(A>T) variant and PCa diagnosis. This variant has a per-allele effect on PCa risk that is similar in magnitude to RPVs in PCa predisposition genes, and should be considered for inclusion in PCa genetic testing panels.

Three-monthly gonadotropin-releasing hormone agonist for ovarian function suppression in premenopausal breast cancer: A systematic review and meta-analysis.

542 Background: Ovarian function suppression (OFS) with gonadotropin-releasing hormone agonists (GnRHa) is a cornerstone of endocrine therapy (ET) for premenopausal women with hormone receptor–positive (HR+) breast cancer (BC), most commonly administered on a monthly (1M) schedule. Three-monthly (3M) formulations have been proposed to improve convenience and adherence, but comparative efficacy and safety versus 1M regimens remain uncertain. Methods: PubMed, Cochrane, and Embase libraries were systematically searched for randomized clinical trials (RCTs) and non-randomized interventional studies comparing 3M versus 1M GnRHa regimens in premenopausal women with HR+ BC receiving ET. Outcomes included the incidence of ovarian escape (OE), mean estradiol (E2), follicle-stimulating hormone (FSH), and luteinizing hormone (LH) levels, disease-free survival (DFS), progression-free survival (PFS), and adverse events, with a minimum follow-up of 12 weeks. Random-effects models were used to estimate pooled risk ratios (RR) and mean differences (MD) with 95% confidence intervals (CI). Subgroup analyses were performed by estradiol assay type, and a sensitivity analysis restricted to RCTs was conducted. Results: Fifteen studies comprising 4,324 patients were included; 2,098 (48%) received a 3M regimen. The overall incidence of OE was low (10%) and did not differ between 3M and 1M schedules (RR 0.86; 95% CI 0.60–1.22). Ultrasensitive assays revealed higher absolute OE rates (31%), without inter-regimen differences (RR 0.90; 95% CI 0.38–2.14). Mean E2 levels at 12 weeks were comparable across dosing strategies (MD 1.36 pg/mL; 95% CI −3.66 to 6.38), with no significant differences in mean FSH or LH levels; analyses restricted to RCTs yielded consistent findings (MD −1.77 pg/mL; 95% CI −5.71 to 2.17). DFS (RR 1.02; 95% CI 0.67–1.56) and PFS (RR 0.86; 95% CI 0.72–1.04) were similar between regimens. The most frequently reported adverse events were hot flashes, arthralgia, headache, and nausea, with no clinically relevant differences between dosing schedules. Conclusions: 3M GnRHa achieved OFS comparable to 1M dosing, with similar survival outcomes and safety profiles in premenopausal women with HR+ BC.

Sacituzumab govitecan in combination with capecitabine for the treatment of advanced gastrointestinal cancers after progression on standard therapy.

TPS4271 Background: Gastrointestinal (GI) malignancies are associated with poor survival and novel therapies are urgently needed. Sacituzumab govitecan (SG) is an antibody-drug conjugate (ADC) that targets the tumor-associated antigen Trop-2 (trophoblast cell-surface antigen 2). SG is covalently bound to the topoisomerase I inhibitor SN-38, which is the active metabolite of irinotecan. Trop-2 is overexpressed in most GI malignancies, including colorectal, gastric, and pancreatic cancers, and is associated with poor prognosis. SG has been approved for the treatment of second-line triple-negative breast cancer and pre-treated hormone-receptor positive metastatic breast cancer but its role in GI cancers is not determined. Furthermore, its combination with capecitabine has never been studied. Methods: We are conducting a single arm, single institution, dose escalation phase 1 trial with a 3+3 design of combination SG plus capecitabine. SG dosing will be at three levels: Level -1 at 5mg/kg; Level 0 at 7.5mg/kg; and Level +1 at 10mg/kg. Dosing will start at dose level 0. SG will be administered as an intravenous infusion on Days 1 and 8 of a 21-day cycle. Capecitabine dose is fixed at 825mg/m 2 and will be delivered orally twice a day for 14 days on and 7 days off, starting on Day 1, of the 21-day cycle. Key eligibility criteria include histologically documented metastatic adenocarcinoma of GI origin, including gastroesophageal, colorectal, and pancreaticobiliary, that has failed standard therapy; age ≥18years; ECOG performance status 0-1; and adequate end organ function. Key exclusion criteria include previous receipt of topoisomerase 1 inhibitors. The primary endpoint is the Recommended Phase 2 Dose (RP2D). Secondary endpoints include objective response rates, duration of response, progression-free and overall survival. An exploratory endpoint is the correlation between Trop-2 expression in collected archival tissue and clinical outcomes. Optional biopsies will be offered for patients with missing or insufficient archival tissue. This study has been registered under NCT06065371 and is actively enrolling. The trial will enroll up to 20 patients. The trial is funded by Gilead Sciences, Inc. Clinical trial information: NCT06065371 .

Neoadjuvant fractionated stereotactic radiotherapy followed by surgical resection for brain metastases: A multicenter, single arm phase II trial (NEO-TACTICS).

2003 Background: Surgical resection followed by adjuvant radiotherapy for brain metastases (BM) is associated with unresolved challenges, including local recurrence (LR), leptomeningeal dissemination (LMD), radiation necrosis (RN), and cognitive decline. Preoperative stereotactic radiosurgery followed by surgical resection has been proposed as a novel strategy to address these limitations. This phase II trial evaluated the efficacy and safety of neoadjuvant fractionated stereotactic radiotherapy (FSRT) followed by surgical resection for BM. Methods: Patients with one index brain metastasis requiring surgical resection (2-5cm diameter) and up to three non-index metastases not requiring resection (< 2cm) were enrolled across 14 centers. All patients received neoadjuvant FSRT (30-35 Gy in 5 fraction) targeting the index lesion, followed by surgical resection. The primary endpoint was 6-month cumulative incidence of LR at the surgical site. Secondary endpoints included LMD, LR at 12 months, RN, distant brain failure (DBF), overall survival (OS), intracranial progression-free survival (IC-PFS), neurocognitive outcomes, and treatment-related adverse events. Results: Between June 2022 and August 2024, we enrolled 57 patients with 53 evaluable for response. The median age was 68 years (range 34-79). The median maximum diameter was 3.2 cm (range 2.0-4.9). The median follow-up duration was 11.9 months (range 1.3-14.7). The 6-month cumulative incidence of surgical site LR was 4.3% (95% CI: 0–10.1) (80%CI：0.4-8.1), and 4.0% (95%CI: 0-9.4) using competing risk analysis. LMD was not observed at either 6, 12 months. The 12-month cumulative incidence of LR was 16.4% (95% CI: 6.0–26.8). Symptomatic RN (≥ Gr 2) was not observed at either 6, 12 months, and asymptomatic RN occurred in 5.9% at 6 months and 8.0% at 12 months. DBF rates were 14.0% (95% CI: 4.4–23.6) at 6 months and 18.1% (95% CI: 7.4–28.8) at 12 months. Median OS rate was 86.3% (95%CI: 77.3-96.3) at 6 months, and 76.5% (95%CI: 65.7-89) at 12 months. IC-PFS rates were 72.5% (95% CI: 61.2-85.9) at 6 months and 64.4% (95% CI: 52.5-79.1) at 12 months. Neurocognitive function was largely preserved, with Mini-Mental State Examination (MMSE) decline ≥1 point observed in 15.4% at 6 months and 9.6% at 12 months, and decline ≥3 points in ≤ 4% at both time points. Adverse events from irradiation to surgery were acceptable (grade ≥2: 10.5%; grade ≥3: 5.3% (CTCAE version 5.0)). Conclusions: Neoadjuvant FSRT followed by surgical resection demonstrated excellent early local control, complete suppression of LMD, acceptable toxicity, and favorable cognitive preservation. This approach represents a promising alternative to postoperative FSRT for resectable brain metastases and warrants further comparative investigation. Clinical trial information: jRCT s042220014.

Updated results of benmelstobart plus anlotinib combined with SOX in the first-line treatment of advanced gastric or gastroesophageal junction (G/GEJ) adenocarcinoma with low PD-L1 expression: A single-arm, multicenter phase II clinical trial.

4054 Background: While PD-1 inhibitors plus chemotherapy have become standard first-line therapy for advanced HER2-negative gastric/gastroesophageal junction (G/GEJ) adenocarcinoma, patients with low PD-L1 expression (Combined Positive Score (CPS)＜5) derive limited benefit. Anti-angiogenic agents can modulate the tumor immune microenvironment and synergize with immune checkpoint inhibitors. Anlotinib, a multi-targeted tyrosine kinase inhibitor approved in China, offers a promising combination strategy. This study evaluates the efficacy and safety of benmelstobart (a PD-L1 inhibitor) combined with anlotinib and SOX (S-1 plus oxaliplatin) as first-line therapy in patients with advanced G/GEJ adenocarcinoma and low PD-L1 expression. Methods: Patients with HER2-negative, unresectable, locally advanced, or metastatic G/GEJ adenocarcinomas and PD-L1 CPS＜ 5, who had not received prior systemic therapy were included. They received benmelstobart (1200mg, iv, d1, q3w) combined with anlotinib (10mg, po, d1~14, q3w), oxaliplatin (130mg/m 2 , d1, iv, q3w) and S-1 (40mg, po, bid, d1~14, q3w) for 6 cycles as initial therapy. Maintenance therapy with benmelstobart (1200mg, iv, d1, q3w) plus anlotinib (10mg, po, d1~14, q3w) followed for non-progressive disease until PD or unacceptable toxicity occurred. Tumor responses were evaluated by RECIST 1.1 criteria. The target sample size was 37, with ORR as the primary endpoint, and safety, DCR, DoR, PFS, and 1-year OS rate as secondary endpoints. Results: From June 2023 to June 2025, 37 patients were enrolled. At the data cut-off date (December, 2025), the best overall response indicated that there were 31 PR (83.8%) and 6 SD (16.2%). Therefore, the preliminary ORR was 83.8% (95%CI: 68-93.8), DCR was 100% (95%CI: 90.5-100). The preliminary prognostic result exhibited that the median PFS of the 37 patients was 11.3 months (95%CI: 8.07-14.53). The 1-year OS rate was 91.33% (95%CI: 75.46–97.12). Safety was manageable, common TRAEs＞20% included platelet count decreased (54.1%), white blood cell decreased (32.4%), anemia (24.3%). Conclusions: The combination of benmelstobart, anlotinib, and SOX demonstrated promising efficacy with a manageable safety profile as first-line therapy for advanced G/GEJ adenocarcinoma with low PD-L1 expression. These findings warrant validation in larger cohorts. Clinical trial information: NCT06939452 .

Microbiota prognostic signature in colon cancer.

e15637 Background: Stage II and III colon cancer (CC) poses a significant challenge due to rising global incidence and mortality rates. Despite advancements in screening and treatment, there is a pressing need for reliable prognostic biomarkers. The objective of this study is to determine the prognostic value of the tumor microbiota in CC patients diagnosed with stage II and III treated with adjuvant chemotherapy. Methods: One hundred and fifty-eight stage II-III CC patients from Hospital Universitario La Paz with FFPE samples and clinical data were included in this study. Proteins were extracted from tumor-rich sections, digested, and analyzed by DIA-MS on an Orbitrap Fusion mass spectrometer. Microbiota proteins related to disease-free survival were defined using Kaplan-Meier and Cox regression. Then, a prognostic signature was built with the selected microbiota proteins and a Cox proportional hazard model. These analyses were done using BRB Array Tools (NIH). Multivariate analysis was performed using SPSS IBM v20. Results: One hundred and fifty-eight CRC patients, with a median age of 67 years, 65 (41%) female, 49 (31%) stage II, 109 (63%) stage III, 59 (37%) right colon, 99 (63%) left colon, 127 (80%) treated with CAPOX, and 31 (20%) treated with FOLFOX, were included. After proteomics analysis, two samples were excluded due to a low amount of protein. Proteomics quantified 341 bacterial proteins, 51 after applying quality criteria. Of those fifty-one bacterial proteins, eleven were related to disease free-survival (p<0.05). A prognostic signature composed by three of these microbiota proteins, from Acinetobacter , Prevotellamassilia , and Staphylococcus , was built. This prognostic signature divides CRC patients into low and high-risk groups (p=0.0019, HR=2.53, 95%CI=1.40-4.36). The DFS at 5 years in the low-risk group is 81.74% whereas in the high-risk group is 59.36%. In a multivariate analysis, including TNM stage, CMS, obstruction, perforation, venous, lymphatic and neural invasion, and differentiation grade; TNM stage, perforation, and the microbiota signature showed prognostic value. Therefore, the microbiota signature provides additional prognostic information to clinical data. Conclusions: Abundance of these three microbiota populations seems to be related to disease-free survival and it should be validated in an independent cohort.