Citations

e12615 Background: Breast MRI after completion of neoadjuvant chemotherapy (NAC) has demonstrated high sensitivity and positive predictive value in evaluating response to NAC. It is widely recognized as the most accurate imaging technique for assessing tumor response, outperforming mammography and ultrasound. However, there remains a knowledge gap in its potential as an early predictor of pathological response, when performed at the midpoint of NAC. Here, we investigated the correlation between MRI findings at the midpoint of NAC and subsequent pathological response at definitive surgery. Methods: A retrospective study was performed of patients with breast cancer who received NAC at Mayo Clinic, Florida and had at least 2 breast MRIs, one at midpoint and one either pre- or post-treatment. Patients with prior ipsilateral breast cancer were excluded. Participants were stratified into three molecular subtypes: hormone receptor-positive (HR+), HER2-positive (HER2+), and triple-negative breast cancer (TNBC). All MRIs were re-reviewed by dedicated breast radiologists, with tumor response classified as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD) based on RECIST criteria. Pathological outcomes, including pathologic complete response (pCR) and residual cancer burden (RCB), were extracted from pathology reports and confirmed by a single breast pathologist, as indicated. pCR was defined as the absence of residual invasive tumor in both the breast and axillary lymph nodes. Results: A total of 154 patients met inclusion criteria from 2012 to 2022. Similar to previous studies, the positive predictive value for the post-treatment MRI was high at 82.4%, with 14 of 17 patients with imaging complete responses (iCR) achieving pCR. For the midpoint MRI, the positive predictive value for patients with iCR was 100%, with 8 out of 8 patients achieving pCR. Patients with certain subtypes of breast cancer with at least PR in both breast and axillary lymph nodes at midpoint MRI had numerically higher pCR rate, specifically HER2+ (22/37, 59.5%) and TNBC (8/13, 61.5%) but not HR+ tumors (12/45, 26.7%). In contrast, patients with less than PR in either breast or axillary lymph nodes showed numerically lower rate pCR in TNBC (7/18, 38.9%) and HR+ tumors (3/18, 16.7%). However, 15/23 (65.2%) patients with HER2+ tumors achieved pCR despite having less than PR at midpoint MRI. Conclusions: Midpoint breast MRI may serve as an early predictor of pathological response, particularly in patients with iCR. Patients with less than PR, especially those with HR+ and TNBC, have lower chances of achieving pCR. These findings highlight the importance of considering additional therapies for these subgroups of patients to improve the treatment outcomes, as they are less likely to achieve pCR with the current standard neoadjuvant chemotherapy regimen.

10512 Background: Polygenic risk scores (PRSs) combine information from single-nucleotide polymorphisms (SNPs) across the genome to explain substantial genetic breast cancer (BC) susceptibility. Previous studies have demonstrated that a multiple-ancestry PRS (MA-385) based on 56 ancestry-informative and 329 BC-associated SNPs is accurate for diverse populations and ranks among the most important known factors affecting the risk of BC development. However, medical guidelines call for more evidence before endorsing the clinical use of PRS, including studies to evaluate possible interactions of SNPs with environmental and hormonal risk factors. Here, we use longitudinal outcomes from the UK Biobank (UKB) to explore interactions of MA-385 and the five most informative individual BC-associated SNPs with the widely used Tyrer-Cuzick (TC) risk model and individual TC risk factors. Methods: The study cohort included 197,509 female UKB participants with no history of cancer at the time of study enrollment. We used Cox proportional hazards models to test associations of MA-385, individual BC SNPs, TC, and individual TC risk factors with BC outcomes. Effect modification of MA-385 and individual SNPs by clinical risk factors was evaluated by including interaction terms in the models. All models were adjusted for age at UKB enrollment and 10 principal components. Hazard ratios (HRs) and 95% confidence intervals (CIs) are reported per standard deviation, and significance tests were performed using two-sided p-values based on likelihood-ratio test statistics. Results: After a median follow-up of 11.8 years, 7,419 (3.8%) participants were diagnosed with BC.In a model including both MA-385 and TC, MA-385 was a significantly better predictor of BC development (HR = 1.56; 95% CI 1.53-1.60, p = 3.2 x 10 -329 ) over the TC model (HR = 1.21; 95% CI 1.18-1.23; p = 5.4 x 10 -69 ). We found no evidence of interaction between MA-385 and TC ( p = 0.31) , nor between individual SNPs and TC. After Bonferroni adjustment for multiple testing (102 tests), we found no evidence of interaction between MA-385 or individual SNPs with clinical TC factors. The strongest evidence for interaction was found between a BC SNP on chromosome 2 (rs13387042) and age at UKB enrollment (unadjusted p = 0.002; Bonferroni adjusted p = 0.23). Conclusions: In a longitudinal analysis of UKB, MA-385 was a highly significant predictor of BC risk and substantially improved prediction over TC. In contrast, interactions of MA-385, and individual BC SNPs, with TC and individual clinical factors in the TC model were not statistically significant. Clinical TC factors have minimal, if any, impact on the strength of the association between MA-385 and BC. These results may help to alleviate concerns expressed in guidelines that SNPs interact with environmental or hormonal risk factors.

1015 Background: T-DXd is an antibody-drug conjugate approved for advanced HER2+/low/ultra-low breast cancer and multiple other solid tumors. T-DXd carries a rare but serious risk of ILD (incidence 12-15%), requiring frequent imaging and symptom evaluation. For > grade (G) 2 ILD, guidelines recommend permanent drug discontinuation. For asymptomatic G1 ILD, drug is held with the option for rechallenge (RC) if imaging findings resolve. Limited data exist on outcomes of RC after ILD in diverse real-world patients (pts). Methods: In this multi-center retrospective study, we analyzed pts with T-DXd related ILD treated from 2017-2024. Pts with ILD were identified via chart/ICD code review. Adjudication of T-DXd related ILD was based on treating providers’ assessment and graded via CTCAE v5. We collected pt demographics, T-DXd and steroid dosing, imaging results, and outcomes after RC. Statistical analysis was performed using Wilcoxon rank sum and Fisher’s exact tests. Results: Four centers treated 712 pts with T-DXd, with a 9.1% rate of any grade ILD (n=65). One other center reported only RC data in 18 pts with ILD. In total, 47 pts were RC; 38 after G1 ILD (81%), 9 after G2. Median (med) time to initial ILD was 145 days (d) after 1st dose (interquartile range [IQR] 78-205). Demographics for pts RC are shown in the table. Among 50 pts with G1 ILD, including pts not RC, 28/50 (56%) received steroids for a med of 36d (IQR 27-79). Radiographic improvement was seen at a med of 24d (IQR 19-63) for pts treated with steroids vs 82d (IQR 48-94) without (p<0.01); and a med of 35d (IQR 22-82) for pts RC vs 81d (IQR 68-105) for pts not RC (p=0.01). Among pts with G1 ILD, 38/50 (76%) were RC at a med of 42d (IQR 36-57) from last dose; 23/38 (61%) were dose reduced. After RC, pts remained on T-DXd for a med of 215d (IQR 60-334); 10/38 (26%) developed recurrent ILD (7-G1, 2-G2, 1-G3) at a med of 211d (IQR 47-273) from RC. No statistically significant differences were seen between ILD onset, time to RC, or demographics for pts with recurrent ILD vs not. Of the 9 pts RC after G2 ILD, T-DXd was continued for a med of 129d (IQR 49-171); 2/9 (22%) developed recurrent ILD (1-G2, 1-G3). No G5 toxicity was seen with RC. Conclusions: In this multi-center study, high RC rates were seen after G1 ILD with long duration of clinical benefit. Pts treated with steroids had faster radiographic ILD improvement, highlighting the importance of early steroid use. Among pts RC after G1 ILD, recurrent ILD rates were low, with the majority G1 and no G5 events. Notably, 9 pts with G2 ILD were RC, with a similar rate of recurrent ILD; this must be interpreted cautiously. Our large cohort data further supports the safety of T-DXd RC in diverse real-world settings. RC Pt Characteristics (n=47) n (%) or Median (IQR) Cancer typeBreastGIGyn 43 (91)3 (6)1 (2) Age (yrs) 57 (52-68) Prior # therapy lines in the advanced/metastatic setting 3 (1-5) Renal impairment (CrCl < 60 mL/min) 8 (17)

3116 Background: Pralsetinib is an oral tyrosine kinase inhibitor that selectively and potently targets oncogenic RET fusion and mutation proteins. RET fusions or mutations are present in various tumor types. We report the final results from the phase 2 portion of ARROW, a phase 1/2, open-label, multi-cohort, dose-expansion study evaluating the efficacy and safety of pralsetinib (NCT03037385) in patients with RET fusion-positive solid tumors other than non-small-cell lung cancer (NSCLC) and thyroid cancer. Methods: Eligible pts were ≥18 years of age with a pathologically documented, definitively diagnosed advanced solid tumor with an oncogenic RET fusion or mutation, had previously received standard of care appropriate for their tumor type, and were not eligible for any other study groups. Overall response rate (ORR) and safety were primary endpoints of the study. Key secondary endpoints included duration of response (DOR), progression-free survival (PFS), and overall survival (OS). The final database lock was May 20, 2024. Results: Twenty-nine patients were enrolled with 11 different solid tumor histologies. Twenty-six (90%) received prior systemic therapy. Median age was 58 years (range 25-75); 59% were female. Twenty-eight patients were included in the efficacy analysis. ORR (by RECIST) was 46.4% (13/28); 10.7% (3/28) achieved complete response (pancreatic cancer, n=2; cancer of unknown primary, n=1) and 35.7% (10/28) achieved partial response. Median PFS was 7 months (95% CI: 3.9, 12.8). Median DOR was 11.1 months (95% CI: 5.5, 25.1). Median OS was 10.3 months (95% CI: 6.8, 25.2). Twenty-five (86%) patients experienced treatment-related adverse events (TRAEs); 19/29 (66%) reported TRAEs ≥grade 3. The most common TRAEs included increased aspartate aminotransferase (11/29; 38%), increased alanine aminotransferase (10/29; 35%), and anemia (9/29; 31%). Four (13.8%) patients experienced hypertension, and 1 (3.4%) patient had ≥grade 3 hypertension. No new safety risks were identified; AEs remained manageable with supportive care and/or dose modifications. Conclusions: In the phase 2 portion of this trial, responses were observed in many tumor types (Table). Pralsetinib demonstrated robust and durable anti-tumor activity with an ORR of 46.4%. These data validate RET fusions as a tissue-agnostic target with sensitivity to RET inhibition and activity beyond NSCLC and thyroid cancer, further supporting the promising potential of pralsetinib to address the unmet medical need in these patients. Clinical trial information: NCT03037385 . Overall response rate by tumor type. Cancer Type (patient n) ORRn (%) Pancreatic (5) 5 (100) Cancer of unknown primary (1) 1 (100) Neuroendocrine (3) 2 (67) Sarcoma (3) 2 (67) Head and neck (2) 1 (50) Small cell lung (2) 1 (50) Hepatobiliary (4) 1 (25) Colorectal (5), gastric (1), ovarian (1), thymic (1) 0

4558 Background: In the past 7 years, 4 immuno-oncology (IO) based combinations have been approved for mRCC. However, QOL data on these combinations are limited to trials in which collection and reporting was not standardized, which further limits cross trial comparisons. Real-world QOL data with multiple treatment regimens are needed to understand how these regimens are tolerated in practice. Methods: ODYSSEY is a multi-site, prospective observational study of 500 US pts with mRCC. Pts must have mRCC (any histology), no prior ST, and follow up at a PCORnet study site. Exclusion criteria include treatment for cancer except mRCC. The primary objective is to determine patterns of change in QOL and symptom burden of pts with mRCC. Minimally important differences (MID) are 3 points for FKSI-19 total score, 1 point for the FKSI-Disease Related Symptoms (DRS) subscale and 7 points for FACT-G. Results: As of 1/6/25, 392 pts were enrolled of whom 299 were managed with ST. Of pts on ST, 114 were treated with IO-IO, 108 with IO-TKI, 33 with IO alone, 27 with Other, and 18 with TKI alone. Median follow up for all pts is 8.8 months (IQR 2.9, 16.2). IO-IO pts are median age 64 yrs, 81% male, 84% white, 56% prior nephrectomy, 84% clear cell; IO-TKI pts median age 66 yrs, 76% male, 92% white, 43% prior nephrectomy, 66% clear cell; IMDC risk profiles are similar. IO-IO pts are more likely to be KPS 100; whereas, IO-TKI pts have a higher median number of metastatic sites and are more likely to have bone or liver metastasis. With median follow-up of 6.0 (IQR 1.8, 14.9) and 8.8 months (4.5, 17.8) in the IO-IO and IO-TKI cohorts, 17 (15%) and 21 (19%) pts had died with a median time to death of 5.4 (2.5, 6.4) and 5.7 months (4.4, 12.8), respectively. 20 (18%) and 16 (15%) pts on IO-IO and IO-TKI had discontinued therapy at a median of 3.0 (IQR 2.0, 4.0) and 6.4 months (2.4, 13.0), respectively. Rates of discontinuation for disease progression and toxicity are similar. Baseline PROs for pts on ST are shown in the Table (higher score indicates better QOL), with RCT data for reference (NA, not assessed). Conclusions: In our prospective multi-center ODYSSEY study, we demonstrate that real world pts treated with contemporary ST have worse baseline QOL than those enrolled on pivotal RCTs. One-third of IO-IO and IO-TKI pts died or discontinued therapy within 6 months of initiation. Our data on real world vs RCT differences in baseline QOL may partially explain the limitations of current IO combination regimens in practice and support development of alternative treatment approaches. Instrument, Mean (SD) ODYSSEYIO-IO(N=114) CheckMate214(N=425) ODYSSEYIO-TKI(N=108) KEYNOTE426(N=402) CheckMate9ER(N=323) CLEAR(N=351) FKSI-19 total 56.0 (12.5) 60.1 (9.8) 53.8 (11.8) NA 58.7 (10.6) NA FKSI-DRS 27.5 (6.3) 30.7 (4.5) 27.0 (5.8) 32 (4.2) 30.2 (5.2) 31.3 (4.4) FACT-G 82.4 (16.7) 82.6 (15.0) 78.5 (16.0) NA NA NA

4553 Background: AS is a recognized strategy in select pts with mRCC to maximize QOL and delay potential toxicity of systemic therapy (ST); however, only 2 prospective studies of AS in mRCC have been published: only 1 with patient-reported outcomes (PRO) and none in the IO-TKI era. Methods: ODYSSEY is a prospective observational study of 500 US pts with mRCC. Eligible pts must have mRCC (any histology), no prior ST, age ≥19. Pts were excluded if treated for non-mRCC cancers or if not followed at a PCORnet study site. Pts completed QOL surveys at baseline, by phone every 3 months for 2 years and then every 6 months until end of follow up. The primary objective is to determine patterns of change in QOL and symptom burden of pts with mRCC. Minimally important differences (MID) are 3 points for FKSI-19 total score, 1 point for the FKSI-Disease Related Symptoms (DRS) subscale and 7 points for FACT-G. Here we report baseline pt characteristics in AS pts compared to ST pts, and baseline QOL differences between these cohorts. Results: As of 1/6/25, 392 pts were enrolled of whom 299 were managed with ST, and 93 pts deferred ST; of these, 53 pts (57%) were classified as AS. Pts on AS are median age 68 yrs, 66% male, 94% white, 81% clear cell, 50% favorable risk, 44% intermediate risk. Compared with ST pts, AS pts were more likely to have undergone nephrectomy (91% vs 53%), favorable risk profile (50% vs 15%), pancreatic metastasis (15% vs 5%), and longer time since RCC diagnosis (median 58 vs 3.3 months); and less likely to have bone, brain, or liver metastasis. After median 8.8 months follow-up (IQR 2.9, 16.2), 2 pts (4%) on AS had died compared with 45 pts (15%) on ST. One pt (2%) on AS started first-line therapy and 45 pts (15%) discontinued ST. Mean baseline QOL (FKSI-19 total, DRS and FACT-G) for AS and ST ODYSSEY pts is shown in the Table (higher score indicates better QOL), with RCT data for reference (NA, not assessed). ODYSSEY pts on AS had higher QOL for all measures compared with ODYSSEY pts on ST. FSKI-DRS was the same or lower for pts on AS compared to the pivotal trials, while ODYSSEY pts on ST had both lower FKSI-19 total and DRS. Conclusions: In our large prospective cohort from ODYSSEY, pts on AS had higher median QOL scores than pts on ST, but similar to those included in RCTs. These results suggest that some RCT pts could have benefitted from AS. Further follow up is needed to determine long term outcomes in pts on AS and how they respond to deferred ST. Clinical trial information: NCT04919122 . Instrument, Mean (SD) ODYSSEY AS (N=53) ODYSSEY ST (N=299) ODYSSEY Difference, AS vs ST (95% CI) CheckMate 214 (N=425) KEYNOTE 426 (N=402) CheckMate 9ER (N=323) CLEAR (N=351) FKSI-19 total 63.4 (9.2) 56.3 (12.5) 7.0(4.0, 10.1) 60.1 (9.8) NA 58.7 (10.6) NA FKSI-DRS 30.8 (4.2) 27.9 (6.0) 2.9(1.5, 4.7) 30.7 (4.5) 32 (4.2) 30.2 (5.2) 31.3 (4.4) FACT-G 87.9 (14.0) 82.4 (16.7) 5.6(1.1, 10.0) 82.6 (15.0) NA NA NA

e18600 Background: Recent studies have suggested that basophilia in myeloproliferative neoplasms (MPN), in general, and in primary myelofibrosis (PMF), in particular, might be associated with poor prognosis. In the current study, we examined for clinical and genotype correlates as well as the prognostic impact of absolute eosinophil (AEC) and basophil (ABC) counts in essential thrombocythemia (ET) and polycythemia vera (PV). Methods: Diagnostic criteria were according to the International Consensus Classification. Study patients were selected based on the availability of information on AEC and ABC. Institutional reference range for AEC was 0.03 to 0.48 x 10 9 /L and for ABC 0.01 to 0.08 x 10 9 /L. "Eosinophilia" was defined as an AEC of ≥0.5 x 10 9 /L. "Basophilia" was evaluated both at an ABC of ≥0.1 x 10 9 /L and ABC of ≥0.3 x 10 9 /L, the latter representing 3 x the upper limit of the normal reference range. Results: Essential thrombocythemia 660 patients were considered. Median (range) values for AEC and ABC were 0.2 x 10 9 /L (0-2.8) and 0.07 x 10 9 /L (0-1.7), respectively. On univariate analysis, AEC and ABC, as continuous variables, were predictive of inferior overall (OS; p<0.01 for both) but not leukemia-free (LFS), myelofibrosis-free (MFFS), venous thrombosis-free (VTFS), or arterial thrombosis-free (ATFS) survival. In univariable analysis, AEC ≥0.5 x 10⁹/L and ABC ≥0.1 x 10⁹/L negatively impacted OS (p<0.01 for both); significance was sustained during multivariate analysis that included IPSET risk variables: AEC ≥0.5 x 10⁹/L (p<0.01; HR 1.9) and ABC ≥0.1 x 10 9 /L (p<0.01; HR 1.7). Polycythemia vera 478 patients were considered. Median (range) values for AEC and ABC were 0.27 x 10 9 /L (0-6.5) and 0.12 x 10 9 /L (0-3.2), respectively. In univariable analysis, neither AEC nor ABC, as continuous variables were predictive of OS, LFS, MFFS, VTFS, or ATFS. AEC ≥0.5 x 10⁹/L also did not have any impact on disease progression or thromboses. However, ABC ≥0.1 x 10 9 /L was associated with a lower risk of post-diagnosis arterial thrombosis (p=0.03; HR 0.6), and ABC ≥0.3 x 10 9 /L was predictive of inferior LFS (p<0.01; HR 6.5) and MFFS (p<0.01; HR 2.5). ABC ≥0.3 x 10 9 /L sustained prognostic relevance in LFS (p<0.01; HR 7.6) during multivariable analysis that included leukocytosis, platelet count, basophil percentage, age, and palpable splenomegaly. Similar findings were noted with regards to MFFS for ABC ≥0.3 x 10 9 /L (p=0.01; HR 2.3) when compared against palpable splenomegaly and age. Conclusions: The current study suggests that AEC ≥0.5 x 10 9 /L and ABC ≥0.1 x 10 9 /L in ET are independently associated with shortened overall survival, while basophilia (ABC ≥0.3 x 10 9 /L) in PV is an independent predictor of leukemic and myelofibrotic disease progression.

3011 Background: BMS-986504 selectively binds to the PRMT5-MTA complex, which represents a synthetic lethal target in MTAP -del cancer cells, while sparing MTAP– wild-type cells. In the first-in-human phase 1/2 CA240-0007 study in advanced, unresectable or metastatic solid tumors with homozygous MTAP -del, BMS-986504 was found to be well tolerated and demonstrated antitumor activity in multiple tumors. Here, we report clinical results and the first PK and PD analyses of BMS-986504 from the dose escalation and expansion phases of CA240-0007. Methods: Pts with measurable/evaluable disease and no available treatment (Tx) with curative intent were enrolled; 7 doses were evaluated (50 to 800 mg) in dose escalation. Objective response rate (ORR), disease control rate (DCR), duration of response (DOR), time to response (TTR), safety, PK, PD, including plasma SDMA were assessed. Results: As of 2 Dec 24, 152 heavily pretreated pts were enrolled across all doses: NSCLC (n = 34), PDAC (n = 41), cholangiocarcinoma (n = 12), and mesothelioma (n = 12) were the most common tumor types. With a median f/u of 9.0 mo (95% CI 7.6–9.9), continued durable antitumor activity and deepening tumor regression were seen across tumor types and doses (ORR = 23%, DCR = 70%, median DOR = 10.5 mo, TTR = 4.6 mo). No new safety signals were identified. Most Tx-related adverse events (TRAEs) were grade (Gr) 1 or 2; 13% had Gr ≥ 3 TRAEs (Gr 3 = 12%, Gr 4 = < 1%, Gr 5 = 0). Doses from 50 to 600 mg QD were assessed for PK/PD. The AUC (0-24) after multiple doses was approximately dose proportional at 200 to 600 mg, and the terminal t 1/2 after a single dose was approximately 24 h (table). There were dose-dependent reductions in predicted plasma SDMA, with the 400 and 600 mg doses approaching the plateau. Conclusions: With longer f/u, BMS-986504 continued to show increasingly durable antitumor activity. BMS-986504 demonstrated a favorable PK/PD profile, supporting QD dosing at 400 and 600 mg. These results support further investigation of BMS-986504 at 400 and 600 mg QD as a potential first-in-class synthetic lethal Tx option in pts with advanced solid tumors with MTAP -del. Clinical trial information: NCT05245500 . PK and PD. 50 mg QD 100 mg QD 200 mg QD 400 mg QD 600 mg QD PK after multiple doses t max a (min–max), h 2.0 (2.0–4.0)n = 3 2.0 (1.0–2.0)n = 3 2.0 (0.5–6.0)n = 8 2.0 (0.5–4.0)n = 10 2.0 (1.0–6.0)n = 10 C max , b ng/mL 107n = 3 377n = 3 685n = 8 1240n = 10 2110n = 10 AUC (0-24) , b h∙ng/mL 948 n = 3 3060 n = 3 7150 n = 7 11,800n = 10 26,700n = 9 AUC (0-24) , b,c h∙ng/mL/mg 19.0n = 3 30.6n = 3 35.8n = 7 29.5n = 10 44.6n = 9 Terminal t 1/2 after single dose, d h 21.7 n = 1 80.7 n = 1 21.5 n = 14 21.7 n = 14 24.1 n = 1 PD Translational exposure targets (C avgss ), X 0.08 0.21 0.5 1.1 1.7 Predicted plasma SDMA reduction, a % (90% PI) 30.0 (23.2–37.8) 38.8 (27.1–47.5) 48.3 (42.0–53.0) 55.0 (50.7–57.6) 57.4 (54.3–59.0) a Median. b Geometric mean. c Dose-normalized. d Arithmetic mean.

Abstract Nearly half of premature deaths can be prevented with healthy lifestyle behaviours. However, most older adults do not adhere to established guidelines to promote a healthy lifestyle. We propose a tailored framework of lifestyle medicine that acknowledges the complex contexts and care needs of older adults. The proposed framework of lifestyle medicine for older adults integrates their lived experiences (e.g. social determinants of health, built environment), the 5 M framework of geriatric medicine (i.e. Mind, Mobility, Medications, Multicomplexity, Matters Most) and the six pillars of lifestyle medicine (i.e. nutrition/diet, physical activity/exercise, restorative sleep, stress management, avoidance of risky substances, positive social connections) through a novel transdisciplinary approach. Guided by the updated framework, key gaps and opportunities are presented to advance lifestyle medicine practice, research and policy with the goal of promoting sustained behaviour change and improving the healthspan of older adults.

AbstractObjectiveEpilepsy is a debilitating disorder affecting more than 50 million people worldwide, and one third of patients continue to have seizures despite maximal medical management. If patients' seizures localize to a discrete brain region, termed a seizure onset zone, resection may be curative. Localization is often confirmed with stereotactic electroencephalography; however, this may require patients to stay in the hospital for weeks to capture spontaneous seizures. Automated localization of seizure onset zones could therefore improve presurgical evaluation and decrease morbidity.MethodsUsing more than 1 000 000 interictal stereotactic electroencephalography segments collected from 78 patients, we performed five‐fold cross‐validation and testing on a multichannel, multiscale, one‐dimensional convolutional neural network to classify seizure onset zones.ResultsAcross held‐out test sets, our models achieved a seizure onset zone classification sensitivity of .702 (95% confidence interval [CI] = .549–.805), specificity of .741 (95% CI = .652–.835), and accuracy of .738 (95% CI = .687–.795), which was significantly better than models trained on random labels. The models performed well across the entire brain, with top five region performance demonstrating accuracies between 70.0% and 88.4%. When split by outcomes, the models performed significantly better on patients with favorable Engel outcomes after resection or who were responsive neurostimulation responders. Finally, SHAP (Shapley Additive Explanation) value analysis on median‐normalized input data assigned consistently high feature importance to interictal spikes and large deflections, whereas similar analyses on histogram‐equalized data revealed differences in feature importance assignments to low‐amplitude segments.SignificanceThis work serves as evidence that deep learning on brief interictal intracranial data can classify seizure onset zones across the brain. Furthermore, our findings corroborate current understandings of interictal epileptiform discharges and may help uncover novel interictal morphologies. Clinical application of our models may reduce dependence on recorded seizures for localization and shorten presurgical evaluation time for drug‐resistant epilepsy patients, reducing patient morbidity and hospital costs.