Real-world Cohort Research Articles

Drug Use Patterns in Myasthenia Gravis: A Real-World Population-Based Cohort Study in Italy.

Background: In the context of a comparative study of efficacy and safety of drugs used in rare neuromuscular and neurodegenerative diseases (CAESAR-call AIFA_FV_2012-13-14), we assessed the use patterns of drugs indicated for myasthenia gravis (MG). Methods: A retrospective cohort study was conducted based on administrative healthcare data. For a cohort of MG patients, prevalent and incident use of pyridostigmine (Py) and other indicated drugs in the first year after case identification was evaluated. Prevalent combined use of major therapies (azathioprine (Az), prednisone (Pr), vitamin D (Vd)) stratified by Py use was assessed, and a comparison between therapies at the time of MG identification and during the first year of follow-up was performed. Results: We included 2369 MG patients between 2013 and 2019. Among them, prevalent and incident Py users were 38.4% and 22.0%, respectively. In the first year of follow-up, the use of Pr was observed in 74.5% of Py prevalent users and in 82.0% of Py incident users, respectively; the use of Az was observed in 24.9% and 23.0%, respectively; and the use of Vd was observed in 53.3% and 48.2%, respectively. Among 910 Py prevalent users, 13.1% also used Az, Pr, and Vd, while 15.3% used none of these. Among 938 non-Py users, 2.7% used Az, Pr, and Vd, while 53.8% used none of these. During the first year, an increase in combined therapies was evident in incident Py users. Conclusions: Our results suggest that, for some MG patients, there may be a need for treatments that combine a rapid onset of benefit with long-term and consistent disease control. These issues may be addressed by the new treatments currently being developed. To date, more studies are needed to address the heterogeneity, quality, and generalizability of the existing data and to evaluate patterns of use, efficacy, and safety of new or emerging therapies for MG.

Validation of a clinical breast cancer risk assessment tool combining a polygenic score for all ancestries with traditional risk factors

PurposeWe previously described a combined risk score (CRS) that integrates a multiple-ancestry polygenic risk score (MA-PRS) with the Tyrer-Cuzick (TC) model to assess breast cancer (BC) risk. Here, we present a longitudinal validation of CRS in a real-world cohort. MethodsThis study included 130,058 patients referred for hereditary cancer genetic testing and negative for germline pathogenic variants in BC-associated genes. Data were obtained by linking genetic test results to medical claims (median follow-up 12.1 months). CRS calibration was evaluated by the ratio of observed to expected BCs. ResultsThree hundred forty BCs were observed over 148,349 patient-years. CRS was well-calibrated and demonstrated superior calibration compared with TC in high-risk deciles. MA-PRS alone had greater discriminatory accuracy than TC, and CRS had approximately 2-fold greater discriminatory accuracy than MA-PRS or TC. Among those classified as high risk by TC, 32.6% were low risk by CRS, and of those classified as low risk by TC, 4.3% were high risk by CRS. In cases where CRS and TC classifications disagreed, CRS was more accurate in predicting incident BC. ConclusionCRS was well-calibrated and significantly improved BC risk stratification. Short-term follow-up suggests that clinical implementation of CRS should improve outcomes for patients of all ancestries through personalized risk-based screening and prevention.

Characterization of mutational signatures demonstrating adaptive mutability post anti-EGFR therapy in a real-world metastatic colorectal cancer cohort.

3545 Background: Clinical benefit from anti-EGFR therapy (EGFRi) for RAS/RAF WT metastatic colorectal cancer (mCRC) is limited by secondary mutations in mitogen-activated protein kinases. The source of these new clones has been debated. In contrast to the outgrowth of pre-existing rare subclones, a new model, termed adaptive mutability, proposes the transient loss of DNA repair genes and concurrent induction of error-prone polymerase repair mechanisms (Russo et al Science, ’19). However, clinical evidence for this effect in a large patient cohort is lacking. Using longitudinal circulating tumor DNA (ctDNA) in a real-world cohort, we sought to explore mutational signatures and assess the presence of adaptive mutability in mCRC patients after EGFRi treatment. Methods: 288 patients from Guardant Health’s GuardantINFORM database with mCRC who received EGFRi and underwent longitudinal ctDNA NGS testing by Guardant360 (Redwood City, CA) between December 2014 and March 2023 were included. We extracted de novo single-nucleotide variant mutational signatures from the somatic ctDNA mutations per-cohort level, comparing pre- to post-treatment mutational profiles. Extracted mutational signatures were compared against COSMIC catalogue. Results: Of the 288 patients identified, 99.7% (287/288) were MSS/MSI-L prior to treatment. Pre-treatment, patients had a mean and median of 3 reportable variants per sample, with 30.2% (87/288) samples with >5 variants. Post-treatment, patients had a mean of 5 and a median of 4 reportable variants, with 64.6% (186/288) of patients with >5 variants. Overall mean shift in reportable variants between pre- and post-EGFRi was 5 variants, with a median of 4 changes. In comparing pre- and post-EGFRi cohorts for new mutations, three mutational signatures emerged: SBS17b, SBS15, and SBS3. SBS15 is associated with mismatch repair (MMR) deficiency and classically seen in MSI-H tumors, though notably this cohort was overwhelmingly MSS. SBS15 has been associated with polymerase epsilon exonuclease (POLE) defects and likely reflects repair by alternate error-prone polymerases such as Polι, Polλ, and Polκ. SBS3 has been associated with homologous recombination (HR) deficiency. SBS17b has been associated with exposure to 5-FU based chemotherapy in CRC patients, which was co-administered in many of the EGFRi treated patients. Conclusions: New mutations arising with EGFR inhibition demonstrate strong evidence for adaptive mutability, showcasing loss of mismatch repair and HR fidelity, and reliance on error-prone polymerases. This is the first large cohort to indirectly demonstrate evidence for this genomic mechanism underlying secondary acquired resistance. Further validation of the adaptive mutability signatures may support therapeutic strategies leveraging this vulnerability to extend benefit from targeted therapies.

Improving lung cancer health equity by applying deep learning to low dose CT screening of minority and disadvantaged patients.

1584 Background: In the US, disparities in lung cancer mortality exist for African American, Hispanic, and other minorities. Standard of care low dose CT screening (LDCT) detects early-stage disease and improves mortality, yet disparities are perpetuated in screening by eligibility criteria derived from cohorts underrepresenting these minorities. One such cohort is the National Lung Screening Trial (NLST) cohort which is 92% White. Consequently, guidelines for lung cancer screening may be insufficient to address the unique needs of diverse populations. We hypothesize that Artificial Intelligence prediction of future lung cancer risk from an individual’s LDCT can partially mitigate racial and ethnic disparities and improve health system practice guidelines by individualizing screening risk as compared to current general guidelines. Here, we benchmark a Resnet18 3D neural network trained on NLST LDCT images, Sybil, on the diverse patient population of the University of Illinois Health system (UIH) which is 20% White and 60% African American. Methods: A real-world cohort from UIH consisting of 1,450 CT studies was evaluated alongside 60,378 CT studies from the NLST cohort. All CT studies evaluated by the model were not used in model training. Using Youden’s J index as a probability cutoff, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were evaluated. Receiver operating characteristic (ROC) and precision-recall (PR) curves were generated to assess model performance between cohorts. NLST data were truncated to achieve equivalent incidence of lung cancer with UIH when generated PR curves. Results: Multi-year prediction performance (ROC-AUC and PR-AUC) between cohorts are summarized (Table). For prediction of lung cancer within 1-year of LDCT in the UIH cohort, the model respectively demonstrated sensitivity, specificity, positive predictive value, and negative predictive value among White (0.80, 0.77, 0.54, 0.92), African American (0.84, 0.78, 0.39, 0.97) races and Hispanic (0.75, 0.73, 0.60, 0.84) and non-Hispanic (0.87, 0.77, 0.42, 0.97) ethnicities. Conclusions: Model performance was similar between the NLST (92% White) and a diverse, real-world cohort at UIH (20% White) though decreases in ROC-AUC performance in year 1 predictions and may be due to insufficient representation of minority populations during model training. Prospective studies involving larger and more representative patient populations should be conducted to further optimize the model and evaluate its clinical utility to improve lung cancer health equity in minority populations.[Table: see text]

Outcomes in patients (pts) with relapsed/refractory (R/R) diffuse large B cell lymphoma (DLBCL) treated with systemic therapy from real world experience (ORCHID) study.

e19067 Background: Odronextamab, an off-the-shelf CD20×CD3 bispecific antibody, has shown compelling efficacy (objective response rate [ORR] 52%, complete response [CR] rate 31.5%) and a generally manageable safety profile in pts with heavily pretreated R/R DLBCL in the single-arm ELM-2 trial (NCT03888105; Ayyappan et al. ASH 2023). We evaluated outcomes in pts with R/R DLBCL treated with currently available third-line or later (3L+) therapies in a real-world setting (ORCHID; NCT05338892) to better evaluate results from ELM-2 in the absence of a randomized control arm. Methods: ORCHID is a multicenter, retrospective, observational study, using electronic medical record or research databases, of pts with R/R DLBCL who have received ≥2 prior lines of therapy (LoTs; including an anti-CD20 and an alkylator) and initiated 3L+ systemic therapy for DLBCL between January 1, 2015 and June 30, 2021. After applying similar eligibility criteria to those in the ELM-2 trial, inverse-probability of treatment weighting (IPTW) was used to balance key prognostic variables between the trial and real-world cohorts. Pts were followed from the start of qualifying LoT until death, end of study period (December 31, 2021), or loss to follow up, whichever occurred first. The primary endpoint was ORR per Lugano criteria assessed by independent central review (ICR). Secondary endpoints included ICR-assessed CR rate, disease control rate (DCR), progression-free survival (PFS), duration of response (DOR), and overall survival (OS) (see Thieblemont et al. ASH 2022). Results: Pts in ORCHID (N=148) generally had less severe disease than trial pts prior to IPTW. The most commonly used regimens were anti-CD20 based therapies, CAR T, and chemotherapy alone. The IPTW cohort had a median age 65 y, 79% Ann Arbor stage III/IV, 66% primary refractory, 88% refractory to last LoT, and median (range) prior LoT: 2 (2–6). In the IPTW cohort, ORR was 44%, CR 18%, and DCR 50%. Median (m) PFS (mPFS), DOR (mDOR), and OS (mOS) were 7.2, 9.0, and 10.8 mo, respectively. Given the lack of regular scans in routine practice (47% of pts had only one ICR response assessment over the study period), sensitivity analysis that considered subsequent treatment as progression resulted in shorter mPFS (4.6 mo) and mDOR (6.1 mo). In the majority of LoTs initiating non-CAR T therapy (n=111), ORR was 33%, CR 15%, DCR 43%. The mOS in non-CAR T group was 7.5 mo and was not reached in those intended for CAR T. Conclusions: Pts in the ORCHID study, especially those who did not receive CAR Ts, had markedly lower response rates than those in the R/R DLBCL cohort of ELM-2. Despite prior failure, most pts still received an anti-CD20 based regimen in 3L+ setting in routine practice, highlighting the need for treatment options with different mechanisms of action that provide deep and durable responses and improve outcomes in this pt population. Clinical trial information: NCT05338892 .

Patterns and risk factors of immune-related adverse events in a real-world cohort with lung cancer receiving immunotherapy.

12130 Background: Immune checkpoint inhibitors (ICIs) have emerged as a core pillar of lung cancer (LC) therapy, but ICIs are commonly associated with a spectrum of immune-related adverse effects (irAEs). The real-world patterns and risk factors of irAEs in LC remain uncertain. Methods: Patients with LC newly started on ICIs between 10/01/2018 and 09/30/2021 were retrospectively collected from the Tufts Medical Center cancer registry and pharmacy records. irAEs occurring within 12 months after initiation of ICIs were identified. Univariable logistic regression was used to assess the risk factors of irAEs. Those with p-value<0.20 were included in multivariable logistic regression analysis (MVA), along with clinically relevant factors. Results: Of 125 LC adult patients (median age: 70 years, 68 males), 104 had non-small cell lung cancer (NSCLC), 123 had advanced stages, and 16 had preexisting autoimmune diseases (ADs). Six patients were treated with dual-agent ICIs. Pembrolizumab was the most often used single agent (67.2%). In total 50 irAEs occurred in 39 patients. The most common irAEs were endocrinopathies (34%), pneumonitis (20%), dermatitis (14%), and gastrointestinal toxicity (14%) with the median onset time of 148, 136, 19, and 68 days, respectively. 70% of irAEs were grade 1 or 2. 44% of irAEs were treated with immunosuppression, 38% were referred to specialist care, and 30% required hospitalizations. 56% of irAEs resolved within 12 months and 67.9 % of those were rechallenged with ICIs. In exploratory univariable analyses among all the patients, age, small cell lung cancer (SCLC), PDL1 positivity (TPS≥1%), concurrent NSAIDs use, and radiation therapy had p< 0.2. In MVA, SCLC remained significantly associated with irAEs (OR=4.14, 95% CI [1.50, 12.09], p=0.007). Among NSCLC subset (N=104), age, PDL1 positivity, concurrent NSAIDs, and acetaminophen use had p<0.2. Age (OR=0.94, 95% CI [0.883, 0.990], p=0.026) and PDL1 positivity (OR=3.17, 95% CI [1.12, 9.87], p=0.036) remained significant in MVA (Table). Conclusions: Our study described the real-world patterns of irAEs in LC patients. SCLC was found to be an independent risk factor of irAEs in our cohort. In NSCLC, younger age and PDL1 positivity were associated with irAEs occurrence. Future studies are required to validate these findings in larger samples. [Table: see text]

Dual immune checkpoint inhibition with nivolumab and ipilimumab (N+I) in patients with advanced anaplastic thyroid carcinoma (ATC).

e18075 Background: There are no standard of care therapies in advanced incurable setting for BRAF wild type (WT) ATC, & after dabrafenib and trametinib (D+T) for BRAF V600E mutant ATC. We previously showed promising clinical activity of dual inhibition with N+I in exploratory cohort of 10 patients (pts) with advanced ATC in a single arm phase II clinical trial, with objective response rate (ORR) of 30% (NCT03246958, Lorch ASCO 2020). Here, we present efficacy and safety outcomes from a real-world cohort of pts with advanced ATC treated with N+I regimen. Methods: We conducted a retrospective cohort study of pts with advanced ATC (unresectable or with distant metastases, DM) who were seen at the Dana-Farber Cancer Institute (DFCI) thyroid cancer center from July 2021 to Jan 2024, received at least one dose of N+I with palliative intent & completed at least one response assessment (data cutoff, 1/31/24). Pts on the DFCI clinical trial were not included. Investigator-assessed ORR (for the first N+I based treatment regimen) and immune-related adverse events (irAEs, collected till cutoff regardless of ongoing treatment) were evaluated. Results: 19 pts with advanced ATC (5: unresectable disease, 14: DM) were included, BRAF WT 12/19 (63%). Median age was 70y (range, 34-81), 68% female. ECOG PS at start was 0-1 in 14/19 (74%), 2 in 4 (21%) & 3 in 1 (5%) pts. 5/19 (26%) pts received N+I concurrently with D+T; 4/5 (80%) in 2nd line after either mixed response or progressive disease (PD) on D+T. Best overall response (BOR) in N+I+D+T group was partial response (PR, 1/5, 20%) & stable disease (SD, 4/5, 80%); ORR 20%. 14/19 (74%) pts received N+I, 12/14 (86%) as 1st line palliative therapy. 5 pts in N+I group (all with DM) received local therapies concurrently (1: debulking thyroid surgery, 3: radiation therapy (RT) to thyroid, & 1: palliative RT to bone DM). BOR in N+I group were: complete response, CR (4/14, 28.6%); PR (3, 21.4%); SD (4, 28.6%); & PD (3, 21.4%); ORR 50%. Excluding 5 pts with concurrent local therapies, ORR was 33.3% (1/9 CR, 2/9 PR).PD-L1 TPS was available for 12/19 pts (4: N+I+D+T, 8: N+I) and ≥1 in all tumors (range, 1-100). In 8 pts in N+I group, ORR stratified by PD-L1 TPS: TPS <50 (range, 1-30): 0% (0/2: 1 SD, 1 PD), & TPS ≥50 (range, 80-100): 83.3% (5/6: 3 CR, 2 PR, 1 PD). Any grade irAEs were observed in 16/19 (84.2%) pts. ≥Grade 3 irAEs occurred in 11/19 (57.9%) pts, most frequent were colitis (5, 26.3%), dermatitis/pruritus (3, 15.8%), myocarditis (2, 10.5%), & adrenal insufficiency (2, 10.5%). There were no treatment-related deaths. Conclusions: In this largest real-world cohort study of pts with advanced ATC, N+I therapy showed high ORR of 50% (congruent with DFCI trial), albeit with high rates of iRAEs. High ORR was seen with PD-L1 TPS ≥50. Survival outcomes and genomic biomarkers, currently under investigation, will be presented at the meeting. N+I merit further investigation in an ATC focused clinical trial.

Clinical characteristic and survival outcomes of patients with advanced NSCLC according to KRAS mutational status in the French real-life ESME cohort

The RAS/MEK signaling pathway is essential in carcinogenesis and frequently altered in non-small-cell lung cancer (NSCLC), notably by KRAS mutations (KRASm) that affect 25%-30% of non-squamous NSCLC. This study aims to explore the impact of KRASm subtypes on disease phenotype and survival outcomes. We conducted a retrospective analysis of the French Epidemiological Strategy and Medical Economics database for advanced or metastatic lung cancer from 2011 to 2021. Patient demographics, histology, KRASm status, treatment strategies, and outcomes were assessed. Of 10 177 assessable patients for KRAS status, 17.6% had KRAS p.G12C mutation, 22.6% had KRAS non-p.G12C mutation, and 59.8% were KRASwt. KRASm patients were more often smokers (96.3%) compared with KRASwt (85.8%). A higher proportion of programmed death-ligand 1 ≥50% was found for KRASm patients: 43.5% versus 38.0% (P < 0.01). KRASm correlated with poorer outcomes. First-line median progression-free survival was shorter in the KRASm than the KRASwt cohort: 4.0 months [95% confidence interval (CI) 3.7-4.3 months] versus 5.1 months (95% CI 4.8-5.3 months), P < 0.001. First-line overall survival was shorter for KRASm than KRASwt patients: 12.6 months (95% CI 11.6-13.6 months) versus 15.4 months (95% CI 14.6-16.2 months), P= 0.012. First-line chemoimmunotherapy offered better overall survival in KRAS p.G12C (48.8 months) compared with KRAS non-p.G12C (24.0 months) and KRASwt (22.5 months) patients. Second-line overall survival with immunotherapy was superior in the KRAS p.G12C subgroup: 12.6 months (95% CI 8.1-18.6 months) compared with 9.4 months (95% CI 8.0-11.4 months) for KRAS non-p.G12C and 9.6 months (8.4-11.0 months) for KRASwt patients. We highlighted distinct clinical profiles and survival outcomes according to KRASm subtypes. Notably KRAS p.G12C mutations may provide increased sensitivity to immunotherapy, suggesting potential therapeutic implications for sequencing or combination of therapies. Further research on the impact of emerging KRAS specific inhibitors are warranted in real-world cohorts.

Liver histology is associated with long-term clinical outcomes in patients with metabolic dysfunction-associated steatohepatitis.

Few studies have examined the risk of long-term clinical outcomes in patients with metabolic dysfunction-associated steatohepatitis in relation to liver histology. We aimed to study this using a real-world cohort. Adults (N = 702) recorded on Vanderbilt University Medical Center's Synthetic Derivative database (1984-2021) with evidence of metabolic dysfunction-associated steatohepatitis on liver biopsy were followed from the first biopsy until the first clinical event or last database entry (median: 4.7y). Risks of cirrhosis (N = 650), other noncirrhotic liver-related (N = 702) and cardiovascular-related outcomes (N = 660), and mortality due to liver, cardiovascular, or cancer events (N = 660) were determined as a function of baseline histology (fibrosis stage [F], lobular inflammation grade [LI], hepatocyte ballooning grade [HB], and steatosis score) adjusting for sex, age, diabetes, and weight-loss surgery. Cirrhosis risk was reduced for lower versus higher fibrosis stage (HR: F0-1 vs. F3: 0.22 [95% CI: 0.12-0.42]), LI1 versus LI2-3 (0.42 [0.19-0.97]), and HB1 versus HB2 (0.20 [0.08-0.50]). Lower fibrosis stage was associated with significantly lower risks of liver-related outcomes versus F4 cirrhosis (eg, F0-1: 0.12 [0.05-0.25]), whereas no differences were seen across baseline lobular inflammation, hepatocyte ballooning, and steatosis grades/scores. Lower versus higher lobular inflammation grade was associated with lower risks for liver-related outcomes in patients with weight-loss surgery. There was a trend for lower risks for cardiovascular-related and any long-term outcomes with lower versus higher fibrosis stage. Fibrosis stage and lobular inflammation and hepatocyte ballooning grades predict the risk of long-term outcomes, supporting the use of these histological features as potential surrogate markers of disease progression or clinical outcomes.

Clinical outcomes and impact of mismatch repair deficiency (dMMR) in patients with pT4N0 colon cancer (CC): Data from a large, consecutive, multicenter, real-world cohort.

3619 Background: T4 is widely recognized as one of the most important and independent prognostic factors in adjuvant setting for CC patients. However, the optimal management in terms of adjuvant treatment in stage II (pT4N0) as well as the impact of DNA mismatch repair deficiency (dMMR) in this subset of patients remain unclear. Here, we present a large, multicenter, real-world analysis of pT4N0 CC patients. Methods: A real-world database including pts treated at 9 Italian institutions from April 2010 to November 2023 was queried. Clinico-pathological characteristics of pts diagnosed with a stage II pT4N0 CC were analyzed. Pts were stratified according to duration of treatment, DNA mismatch repair (MMR) status, type of adjuvant chemotherapy (ACT). The primary endpoints were median relapse-free survival (RFS) and overall survival (OS). Results: A total of 380 pts was included, and outcomes data were available for 288 pts.Median age was 73 years and 49% of pts were male. 47 pts (16%) had dMMR CC, among which Lynch Syndrome was diagnosed in 19%. 10% of pts had BRAF mutant tumor, and 38% KRAS mutant tumor. After a median follow-up of 40.6 months (37.9-49.3), 3-yr RFS and 5-yr RFS were 70% and 55%, respectively, regardless treatment; 3-yr OS was 88% and 5-yr OS 73%. 154 pts (53%) received ACT: oxaliplatin (oxa)-based CT was given to 103 pts (67%), compared to 51 pts who received monotherapy (33%). Six-month ACT was given to 38 pts (74.5%) and 75 pts (72.8%) in the monotherapy and oxa-based group, respectively. Pts receiving 6-month ACT had a significant longer OS compared to those who did not receive ACT (HR 0.17; 0.08-0.34; p &lt;.001). A trend towards improved OS was observed in pts receiving 6-month ACT versus pts receiving 3-month ACT (HR 0.46; 0.16-1.29; p 0.129). Similar results were observed in terms of RFS: 6-month ACT showed a longer RFS compared to 3-month ACT (HR 0.58; 0.32-1.05; p .068), and a significantly longer RFS compared to no-ACT (HR 0.50; 0.32-0.77; p .001). Oxa-based ACT was associated with a significant improvement in OS compared to monotherapy (HR 0.30; 0.11-0.85; p .016). Overall, pts with dMMR CC did not show a statistically significant difference in terms of OS and RFS (p .66 and .19 respectively) compared to proficient MMR pts. Conclusions: In this large real-world dataset, T4 was confirmed being a key poor prognostic factor, independently to MMR status. We observed that ACT provides large benefit in this population, and oxa-based ACT may be better than monotherapy. In addition, 6-month oxa-based ACT showed a trend towards a better outcome in comparison to 3-month ACT, even if not statistically significant. Given the retrospective nature of the study, these data should be validated in larger and prospective cohorts; however, this work may help oncologists in shared decision making in this specific subgroup of patients.

A digital pathology AI model to predict immune-oncology biomarker status and pembrolizumab response in a real-world cohort of patients with colorectal and breast cancer.

e13617 Background: Deep learning (DL) algorithms have garnered significant attention for prediction of biomarker status in digitized tissue slides. Here, we trained and validated a digital pathology algorithm to predict microsatellite instability (MSI)/mismatch repair deficiency (MMRd) in colorectal cancer (CRC) and PD-L1 status in breast cancer (BC) using a large real-world patient cohort. Methods: H&amp;E and immunohistochemistry (IHC) slides of CRC and BC from Caris Life Sciences (Phoenix, AZ) with known biomarker status were digitally scanned (CRC: N=27,494, 6.6% MSI-high by next-generation sequencing [NGS]; N=25,788, 6.4% MMRd by IHC; BC: N=16,069, 15.8% PD-L1+ by IHC). A transformer-based DL model (Wagner, Sophia J., et al., 2023) was trained to predict biomarker status, with 5-fold cross-validation. Accuracy, F1 score, area under the curve (AUC), and other metrics were calculated comparing the model to NGS/IHC “truth”. Risk stratification was performed on holdout datasets of pembrolizumab-treated CRC (N=422-454) and triple negative breast cancer (TNBC; N=513) pts. Post-treatment overall survival (OS) was calculated from start of pembrolizumab to last contact, per insurance claims data. Hazard ratios (HR) and 95% confidence intervals (CI) were calculated by Cox-Proportional hazards and compared between model and “truth”. Results: When applied to holdout datasets for biomarker prediction, the validated model achieved AUC of 0.78-0.95. In CRC and BC, performance was better among primary specimens than metastatic. For MSI-H/MMRd risk stratification in pembrolizumab-treated CRC, HR for OS were comparable between model and “truth” for primary and metastatic cohorts. When applied to the TNBC holdout dataset, the model performed better than “truth” for PD-L1 risk stratification in pembrolizumab-treated pts. Conclusions: A digital pathology DL model predicts biomarker status with high accuracy and is non-inferior to NGS/IHC for risk stratification of pembrolizumab-treated CRC and TNBC pts. These results support further investigation of AI algorithms developed from large datasets as possible rapid screening tools for biomarker detection in molecular testing programs. [Table: see text]

Results from the follicular lymphoma (FL) outcomes in patients (pts) with relapsed/refractory (R/R) disease treated with systemic therapy in a real-world assessment (FLORA) study.

7076 Background: Odronextamab, an off-the-shelf CD20×CD3 bispecific antibody, has shown compelling efficacy (objective response rate [ORR] 80%, complete response [CR] rate 73%) and a generally manageable safety profile in pts with heavily pretreated R/R FL in the single-arm ELM-2 trial (NCT03888105; Villasboas, et al. ASH 2023). We evaluated outcomes in pts with R/R FL treated with currently available third-line or later (3L+) therapies in a real-world setting (FLORA; NCT05338879) to better evaluate the results from ELM-2 in the absence of a randomized control arm. This is the first real-world study in 3L+ FL that evaluated scans from 100% of the cohort. Methods: FLORA is a multicenter, retrospective, observational study, using electronic medical record or research databases, of pts with R/R FL who have received ≥2 prior lines of therapy (LoTs; including an anti-CD20 and an alkylator) and initiated 3L+ systemic therapy for FL between January 1, 2015 and December 31, 2020. After applying similar eligibility criteria to those in the ELM-2 trial, inverse-probability of treatment weighting (IPTW) was used to balance key prognostic characteristics between the trial and the real-world cohorts. Pts were followed from the start of qualifying LoT until death, end of study (December 31, 2021), or lost to follow-up, whichever occurred first. The primary endpoint was ORR assessed by independent central review (ICR). Secondary endpoints included ICR-assessed CR rate, disease control rate (DCR), progression-free survival (PFS), duration of response (DOR), and overall survival (OS) (see Bachy, et al. ASH 2022). Results: Pts in FLORA (N=100) generally had less severe disease than trial pts prior to IPTW. The most commonly used regimens were chemotherapy + anti-CD20, chemotherapy alone, and anti-CD20 + immunomodulatory agent. The IPTW cohort had a median age of 61 y, 59% POD24, 56% high FLIPI, 60% chemo-immuno refractory, 73% refractory to last LoT, and median (range) prior LoT: 3 (2–8). In the IPTW cohort, ORR was 52%, CR 31%, and DCR 62%. Median PFS (mPFS) was 11.5 mo, median OS was 26.1 mo, and median DOR (mDOR; responders in IPTW cohort: n=54) was 28.2 mo. Given the lack of regular scans in routine practice (37% of responders did not have any subsequent ICR assessment after CR/PR), sensitivity analysis that considered subsequent treatment as a progression event resulted in shorter mDOR (8.1 mo) and mPFS (7.2 mo). Conclusions: Pts in the FLORA study had markedly lower response rates than those in the R/R FL cohort of ELM-2. This study also showed that most pts still received an anti-CD20–based regimen in 3L+ in routine practice, highlighting the need for treatment options with different mechanisms of action that provide deep and durable responses and improve outcomes in this pt population. Clinical trial information: NCT05338879 .

467P Novel staging schemes for Siewert type II esophagogastric junction adenocarcinoma: A real-world data cohort study from SEER database

467P Novel staging schemes for Siewert type II esophagogastric junction adenocarcinoma: A real-world data cohort study from SEER database

Effect of tumor burden on survival in patients with stage IV EGFR-mutant NSCLC receiving osimertinib monotherapy.

8611 Background: Osimertinib (osi) has traditionally been given as monotherapy for patients (pts) with stage IV EGFR-mutant (EGFRm) non-small cell lung cancer (NSCLC), although acquired resistance remains a challenge. Recent data from the FLAURA2 study demonstrated a progression free survival (PFS) benefit of approximately 9 months (mos) when osi is combined with platinum-pemetrexed chemotherapy. Despite this improvement, combination therapy is associated with higher rates of adverse events, making it challenging to decide which pts should receive TKI + chemo vs TKI monotherapy. Tumor burden (TB) in pts with stage IV EGFRm NSCLC, which may be associated with risk of developing resistant clones on TKI monotherapy, and its association with clinical outcomes has not been examined. Methods: Retrospective, single institution manual case review of pts who received osi monotherapy. Primary objectives were PFS (i.e., oligoprogression requiring local therapy, disease progression (PD) with osi discontinuation, or death) and overall survival (OS) related to extent of TB. Those who did not receive osi monotherapy for stage IV EGFRm NSCLC or discontinued osi for reasons other than PD were excluded. Demographics, comorbidities, performance status, PD-L1 expression, histology, EGFRm subtype, prior TKI use, total number of visceral tumors (TTn), and total number/sites of metastatic disease (MDn) collected for analysis. Extent of TB categorized by TTn (1-4, 5-9, ≥10) and MDn (1-2, 3-4, ≥5). PFS and OS estimated by Kaplan-Meier method and compared using log-rank test. Results: 325 pts eligible for analysis. Median follow-up time was 22.7 mos. Median PFS (mPFS) for the overall cohort was 16.3 mos (95% CI; 13.7 - 18.9) and median OS (mOS) was 31.1 mos (95% CI; 27.4 - 39.3). Pts with a TTn of 1-4 had significantly better mPFS compared to those with a TTn of 5-9 and ≥10 (mPFS: 21.5 vs 16.3 vs 13.4 mos, respectively; P = 0.001). Pts with a TTn ≥10 had significantly worse mOS compared to those with a TTn of 5-9 and 1-4 (23.7 vs 35.1 vs 57.7 mos, respectively; P &lt; 0.0001). Similarly, pts with a MDn of 1-2 had significantly better mPFS and mOS compared to those with a MDn of 3-4 and ≥5 (mPFS: 21.6 vs 11.4 vs 9.0 mos, respectively [P &lt; 0.0001]; mOS: 44.3 vs 20.8 vs 22.4, respectively [P &lt; 0.0001]). Pts with liver and bone mets had significantly worse mPFS and mOS, whereas those with lung and brain mets had worse mOS without affecting PFS. Conclusions: The extent of baseline TB, categorized by TTn and MDn, impacts both PFS and OS in pts with stage IV EGFRm NSCLC treated with osi monotherapy. Pts with “low-TB” (i.e., TTn 1-4, MDn 1-2) appear to have a survival advantage compared to those with “intermediate- or high-TB” (i.e., TTn 5-9 and ≥10, MDn 3-4 and ≥5) and may derive less benefit from the addition of chemotherapy to osi. Prospective, matched real-world treatment cohorts are warranted to optimize treatment strategies for this patient population.

Different dosing strategies of pembrolizumab in the treatment of NSCLC: A real-world retrospective observational study.

e23264 Background: Immune checkpoint inhibitors lack a dose-response relationship, allowing optimisation of dosing strategies that may reduce drug use. Pembrolizumab received its initial regulatory approval for the treatment of non-small cell lung cancer (NSCLC) at weight-based dosing (2 mg/kg every 3 weeks (Q3W)), as employed in pivotal clinical trials demonstrating its efficacy and safety. This dosing regimen was later replaced by the currently approved flat-dosing (200 mg Q3W or 400 mg Q6W) in all indications, potentially increasing drug use. This retrospective observational study evaluates differences in pembrolizumab use for different dosing strategies in a real-world patient cohort. Methods: Overall, 124 advanced NSCLC patients, treated with first-line pembrolizumab monotherapy at a single institution, were included. During the observation period, pembrolizumab flat-dosing was employed, with possible extensions of treatment administrations. Patient and treatment data were obtained from medical records. Duration of treatment was calculated from day 1 of first cycle to last day of last cycle, excluding treatment delays due to adverse events. Pembrolizumab use was calculated for 3 dosing strategies: i) as actually received, ii) as flat-dosing, iii) as body-weight dosing, capped at 200 mg Q3W. It is presented as median dose Q3W and cummulative dose in the whole cohort. Results: Included patients had a median weight of 73 kg (IQR: 65-84) and received a total of 1600 mg of pembrolizumab (IQR: 800-3600) over 24 weeks (IQR: 12-60). The median pembrolizumab dose was i) 190 mg Q3W (IQR: 172-200) as actually received by patients, and would be ii) 200 mg Q3W (IQR: 200-200) in case of flat dosing, and iii) 146 mg Q3W (IQR: 130-168) in case of body-weight dosing. Overall, total pembrolizumab dose in the whole cohort was i) 283,600 mg actually used, and would be ii) 327,800 mg in flat-dosing, and iii) 251,200 mg body-weight dosing, resulting in a 23.4% reduction of drug use in weight-adjusted vs flat-based dosing. Conclusions: Pembrolizumab dosing adjusted to body weight, as initially approved, would reduce drug use vs the currently approved flat-dosing, reaching nearly a 25% reduction in our cohort of patients.

378P First-line systemic treatment outcomes in a real-world advanced pancreatic cancer cohort

378P First-line systemic treatment outcomes in a real-world advanced pancreatic cancer cohort

Inherited Germline Variants in Urinary Tract Cancer: A Multicenter Whole-Exome Sequencing Analysis and Correlation With Clinical Features and Tumor Genomics.

This study investigates a real-world multicenter cohort of patients with urinary tract cancer (UTC), with primary disease sites including the bladder, urethra, and upper tract, who enrolled for research molecular testing of their germline and tumor. The purpose of this study was to evaluate factors that could affect the likelihood of identifying a clinically actionable germline pathogenic variant (PV). Patients with UTC were identified from 10 cancer institutes of the Oncology Research Information Exchange Network consortium. The data set comprised abstracted clinical data with germline and tumor genomic data, and comparative analyses were conducted. Clinically actionable germline PVs in cancer predisposition genes were identified in 16 (4.5%) of 354 patients. A higher proportion of patients with the urethra and the upper tract as the primary sites of disease had PVs with a prevalence of 11% (5/45), compared with only 3.6% (11/308) in those with the bladder as the primary site of disease (P = .04). There were no significant differences in markers of genomic instability (such as tumor mutational burden, microsatellite instability [MSI], and loss of heterozygosity, copy number, and chromosomal instability) between those with PVs and those without (P > .05). Of the PVs identified, 10 (62%) were in homologous recombination repair (HRR) genes, three (19%) in mismatch repair (MMR) genes, and three (19%) in genes associated with other pathways. Tissue-based assessment of genomic instability, such as MSI, does not reliably indicate germline PV. A comprehensive clinical germline testing approach that includes HRR genes in addition to MMR genes is likely to yield PVs in approximately one of 10 patients with nonbladder primary disease sites such as the upper tract and the urethra.

Real-World Data on Tezepelumab in Patients With Severe Asthma in Germany

Tezepelumab is a novel biologic blocking thymic stromal lymphopoetin, approved for severe asthma irrespective of biomarker levels or phenotype. To characterize a real-world tezepelumab patient cohort and the efficacy among various asthma phenotypes. We performed a retrospective, multicenter study on patients with severe asthma initiating tezepelumab. Clinical response was evaluated at 3 and 6 months. We included 129 patients with an average age of 52.5 ± 13.1 years, 59.7% were female. The majority (86.0%) had increased type 2 (T2) biomarkers, 68.2% an allergic and 31.8% an eosinophilic phenotype. 23.3% of patients were biologic-naive. 22 (18.2%) patients discontinued tezepelumab therapy owing to suspected side effects or insufficient efficacy. At 6 months' follow-up, median reduction in annualized exacerbation rate was-1 [25th percentile; 75% percentile {-2.9; 0.0}], the reduction of oral corticosteroid dose among patients with long-term oral corticosteroid therapy was -5 mg [-10; 0] and the Asthma Control Test (ACT) improved by 2 [0; 5] points. A treatment response according to Biologic Asthma Response Score of 80.8% was demonstrated. There were no significant differences in treatment response between T2-high versus T2-low, early- versus adult-onset and eosinophilic versus non-eosinophilic asthma. Prior treatment with other biologics was associated with inferior treatment response. In this real-life cohort, including a large proportion of patients with history of previous biologic use and encompassing various subgroups, the majority responded to tezepelumab. Our data further suggest a steroid-sparing effect of tezepelumab.

Phenotyping Hepatic Immune-Related Adverse Events in the Setting of Immune Checkpoint Inhibitor Therapy.

We present and validate a rule-based algorithm for the detection of moderate to severe liver-related immune-related adverse events (irAEs) in a real-world patient cohort. The algorithm can be applied to studies of irAEs in large data sets. We developed a set of criteria to define hepatic irAEs. The criteria include: the temporality of elevated laboratory measurements in the first 2-14 weeks of immune checkpoint inhibitor (ICI) treatment, steroid intervention within 2 weeks of the onset of elevated laboratory measurements, and intervention with a duration of at least 2 weeks. These criteria are based on the kinetics of patients who experienced moderate to severe hepatotoxicity (Common Terminology Criteria for Adverse Events grades 2-4). We applied these criteria to a retrospective cohort of 682 patients diagnosed with hepatocellular carcinoma and treated with ICI. All patients were required to have baseline laboratory measurements before and after the initiation of ICI. A set of 63 equally sampled patients were reviewed by two blinded, clinical adjudicators. Disagreements were reviewed and consensus was taken to be the ground truth. Of these, 25 patients with irAEs were identified, 16 were determined to be hepatic irAEs, 36 patients were nonadverse events, and two patients were of indeterminant status. Reviewers agreed in 44 of 63 patients, including 19 patients with irAEs (0.70 concordance, Fleiss' kappa: 0.43). By comparison, the algorithm achieved a sensitivity and specificity of identifying hepatic irAEs of 0.63 and 0.81, respectively, with a test efficiency (percent correctly classified) of 0.78 and outcome-weighted F1 score of 0.74. The algorithm achieves greater concordance with the ground truth than either individual clinical adjudicator for the detection of irAEs.

Prognostic value of PORTEC-3 molecular markers by disease risk in a real-world early endometrial cancer cohort.

Prognostic value of PORTEC-3 molecular markers by disease risk in a real-world early endometrial cancer cohort.