Risk Stratification Research Articles

Combined Self-Reported and Device-Measured Physical Activity Assessment and Disability Incidence in Older Adults

Combined Self-Reported and Device-Measured Physical Activity Assessment and Disability Incidence in Older Adults

Association of clinical characteristics with the use of proteomic testing for lung nodule malignancy risk stratification at a safety-net hospital.

e20036 Background: The NodifyLung CDT and XL2 tests use proteomics to risk stratify lung nodules. This test has emerged as a valuable tool in facilitating clinical decision making in low to moderate risk lung nodules. Nodify CDT testing classifies lung nodules in three risk groups (High, moderate and No Significant Level of Autoantibodies Detected (NSLAD)) to help further guide diagnostic testing. Nodify XL2, used as a “rule out” test categorizes patients into three groups: Indeterminate, Reduced risk, and Likely benign. We sought to identify clinical and demographic variables and their significance with utilization of the NodifyXL2 test. Methods: A retrospective chart analysis was conducted on all patients who underwent Nodify XL2 testing at the Lung Cancer Screening Clinic at Wyckoff Heights Medical Center, a New York City Safety Net Hospital, from 8/2023 through 9/2024. All patients were found to have NSLAD on the Nodify CDT test. All demographic data including age, smoking status was collected and tabulated. In addition, lung nodule clinical characteristics were obtained and tabulated. Statistical analyses were performed using frequency tables and chi-square tests, with a significance level set at p < 0.05 in efforts to determine clinical significance between higher risk nodules and demographic characteristics. Results: There was a total of 91 patients (40 males and 51 females with any average age of 64 +/- 11.4 years). The Likely benign/Reduced risk classification was found to be statistically significant with specific nodule characteristics, including non-spiculated morphology (p < 0.001) and sub-centimeter size (p< 0.001). Patients less than 70 years old were statistically more likely to be classified as Reduced Risk/Likely Benign lesion (p< 0.011). Additionally, history of non-smoking was also statistically associated with Likely benign/Reduced risk category (p < 0.018). Conclusions: Blood based proteomic testing such as NodifyLung has demonstrated the ability to help further classify low to moderate risk pulmonary nodules with clinical significance. The ability to discriminate between higher and lower risk nodules can further aid in clinical decision making. Blood based proteomic testing is an additional tool that may be used in lung cancer screening/nodule clinics. Association of Nodify XL2 results with demographic characteristics. Indeterminate n (%) Reduced Risk/Likely Benign n (%) p Value Age Group < 70 11 (47.8) 52 (76.5) 0.01084 70 or more 12 (52.2) 16 (23.5) Smoking Current/Former 22 (95.7) 49 (72.1) 0.01817 Never smoker 1 (4.3) 19 (27.9) Nodule Size <10mm 2 (8.7) 38 (55.9) 0.000081 10mm or more 21 (91.3) 30 (44.1) Spiculated Yes 19 (82.6) 17 (25.0) <0.0001 No 4 (17.4) 51 (75.0) Statistical test: Chi Square. “n” represents sample size. “(%)” corresponds to row percentages.

Systematic screening for atrial fibrillation with non-invasive devices: a systematic review and meta-analysis.

Systematic screening for atrial fibrillation with non-invasive devices: a systematic review and meta-analysis.

An AI-powered complete blood count panel for high-grade cervical lesion identification: A multicentric retrospective study.

e17535 Background: Early detection and accurate risk stratification of high-grade cervical lesions (HSIL) are crucial for effective treatment and improved patient outcomes, as timely interventions are both cost-effective and successful in preventing progression to cancer. In this retrospective, multicentric study conducted across five laboratories in Brazil, we propose leveraging machine learning (ML) to repurpose complete blood count (CBC) tests as a cost-effective tool to identify women at high risk for HSIL. Methods: We analyzed complete blood count (CBC) tests from 324,291 women aged 25–64 years who underwent Papanicolaou tests or cervical biopsies within six months of their CBC test, collected between January 2004 and February 2024, across five Brazilian laboratories: Fleury, Amais SP, Labs Amais, Novamed, and Lafe. Of these, 768 women (0.24%) were confirmed as cases with biopsies identifying HSIL, while 323,523 were classified as controls based on Papanicolaou tests negative for malignancy. A ridge regression model was trained using data from one laboratory's database, with the remaining five databases used for testing. Results: Statistical analysis revealed that mean corpuscular hemoglobin (MCH), mean corpuscular volume (MCV), corpuscular hemoglobin concentration mean (CHCM), lymphocyte, and lymphocyte-monocyte ratio (LMR) were significantly higher (p<0.05) in women with HSIL, while age, red blood count (RBC) and RDW were significantly lower. Furthermore, using a feature selection methodology based in a decision tree, we incorporated age, MCH, LMR, and RDW in a ridge regression model, achieving an average external validation AUC of 0.70 ± 0.02, sensitivity of 0.74 ± 0.04, specificity of 0.63 ± 0.06, accuracy of 0.63 ± 0.06, balanced accuracy of 0.68 ± 0.04, negative predictive value (NPV) of 1.00 ± 0.00, and positive predictive value (PPV) of 0.05 ± 0.02. This indicates that by selecting 37% of the population for screening, we could identify approximately 74% of women with HSIL. Conclusions: Our AI model, utilizing CBC parameters, demonstrates potential as a pre-screening tool to identify women at elevated risk for HSIL, thereby optimizing the allocation of Papanicolaou or HPV DNA testing resources. This strategy may be particularly advantageous in resource-limited settings, where access to comprehensive screening programs may be constrained. To support its clinical application, external validation across diverse populations is essential to ensure the model's generalizability and effectiveness.

Assessing the accuracy of Seismofit® as an estimate of preoperative maximal oxygen consumption in patients with hepato-pancreato-biliary, colorectal, and gastro-oesophageal cancer.

Assessing the accuracy of Seismofit® as an estimate of preoperative maximal oxygen consumption in patients with hepato-pancreato-biliary, colorectal, and gastro-oesophageal cancer.

Artificial intelligence (XGBoost) in predicting outcomes among CAR-T therapy patients: The impact of malnutrition and comorbidities using the National Inpatient Sample (2020-2022).

2536 Background: Chimeric Antigen Receptor T-cell (CAR-T) therapy has revolutionized hematologic malignancy treatment but remains costly, with limited access and complications like prolonged hospitalization, sepsis, and mortality. Malnutrition, common in cancer patients, worsens these outcomes. Despite AI’s growing role in oncology, its use in risk stratification for malnourished CAR-T recipients is underexplored. This study leverages the National Inpatient Sample (NIS) 2020-2022 to develop AI-driven models predicting length of stay (LOS), mortality, and sepsis, incorporating the Charlson Comorbidity Index and other factors. Methods: Using the NIS database, adult CAR-T therapy patients were identified with ICD-10 codes. Key variables included demographics (age, gender, race/ethnicity, income), clinical factors (Charlson Comorbidity Index, sepsis, admission type), and hospital characteristics (size, teaching status). AI models (XGBoost, Random Forest, Neural networks) were trained on the 2020 dataset and validated on 2020-2022 data. Hyperparameter tuning via grid search was performed to optimize model performance. LOS was modeled as a continuous outcome, while mortality and sepsis were classified as binary outcomes. Data preprocessing included handling missing values, one-hot encoding of categorical variables, and standardizing continuous variables. SHapley Additive exPlanations (SHAP) were used to interpret feature importance. Results: The study analyzed 1,912 CAR-T hospitalizations over three years, with 11.5% identified as malnourished. AI models demonstrated strong predictive performance, with XGBoost (RMSE: 3.5 days, R² = 0.82) for LOS, Random Forest (AUC: 0.91) for mortality, and Neural Networks (AUC: 0.87) for sepsis. Malnutrition significantly worsened outcomes, increasing LOS by 14.2 days (p < 0.001) and mortality risk by 3.2-fold (p < 0.001). Patients with Charlson Comorbidity Index scores ≥3 had 9.8-day longer LOS and 2.9-fold higher mortality risk (p < 0.001). Racial disparities were evident, with Black patients at 25% higher risk of prolonged LOS and Hispanic patients at increased risk of sepsis (p < 0.05). Malnourished patients in non-teaching hospitals with high comorbidity burdens had the worst outcomes, emphasizing the need for targeted interventions in high-risk populations. Conclusions: AI-driven models incorporating malnutrition and Charlson Comorbidity Index accurately predict LOS, mortality, and sepsis in CAR-T patients. Early identification and management of malnutrition and comorbidities, particularly in racially diverse populations, are critical to improving outcomes. Future research should focus on prospective validation and AI integration into clinical workflows to mitigate disparities.

A Radiomic and Clinical Data-Based Risk Model for Malignancy Prediction of Breast BI-RADS 4A Microcalcifications.

A Radiomic and Clinical Data-Based Risk Model for Malignancy Prediction of Breast BI-RADS 4A Microcalcifications.

Genomics of Suicidal Behaviors: What Can We Learn from Polygenic Scores?

Genomics of Suicidal Behaviors: What Can We Learn from Polygenic Scores?

Quick sequential organ failure assessment and Fournier gangrene severity index as predictors for mortality in Fournier gangrene patients: A retrospective cohort study of 153 patients.

Quick sequential organ failure assessment and Fournier gangrene severity index as predictors for mortality in Fournier gangrene patients: A retrospective cohort study of 153 patients.

Gynecologic oncology referral rates of adnexal masses suspicious for ovarian cancer in an academic health system: A cohort study.

5544 Background: Ovarian-Adnexal Reporting Data System (O-RADS) is an international lexicon and risk stratification tool. O-RADS 4 or 5 lesions are complex adnexal masses that a 10-90% risk of malignancy, and national guidelines recommend gynecologic oncology referral. Our objective was to examine patient, clinician, and imaging factors associated with referral to gynecologic oncology for complex adnexal masses. Methods: This retrospective cohort study was exempt from IRB review. We identified all patients with O-RADS 4 or 5 lesions on ultrasound (US) or MRI from July 1, 2020 to December 31, 2023. Our primary outcome was referral to gynecologic oncology. We gathered patient demographic data and ordering clinician characteristics from electronic health records. We performed descriptive statistics and multivariate logistic regression of patient demographics and ordering clinician characteristics associated with gynecologic oncology referral. Results: Our cohort included 373 patients with O-RADS 4 or 5 lesions and no prior gynecologic oncology care. The referral rate to gynecologic oncology was 68%, and referral within 30 days of abnormal imaging was 43%. Time from abnormal imaging to referral ranged from 0 to 407 days (mean 15.3, median 4 days). In multivariate analyses, the likelihood of referral to gynecologic oncology was higher among patients with repeat abnormal imaging compared to those with single instance of abnormal imaging (aOR 20.61, 95%CI 2.63-161.6), O-RADS 5 lesions compared to O-RADS 4 lesions (aOR 9.15, 95%CI 3.47-24.85) and detection on MRI compared to US (aOR 7.79, 95%CI 1.57-38.65). The likelihood of referral to gynecologic oncology was lower among non-white patients (aOR 0.24, 95%CI 0.08-0.76). There were no differences by Hispanic ethnicity, rurality, insurance, or language. Referral was higher among patients whose imaging was ordered by an internal medicine clinician (aOR 3.89, 95%CI 1.48-10.20) compared to ob/gyn. Conclusions: One-third of patients with complex adnexal masses were not referred to gynecologic oncology. Disparities in referral to gynecologic oncology for complex adnexal masses rates based on patient race and ordering clinician specialty highlight the need for system-based approaches including clinician education or automated referrals. Gynecologic oncology referral after O-RADS 4/5. Multivariate OR (95%CI) Postmenopausal (≥55 years) 1.89 (0.95-3.74) Race - White Reference - Black 0.57 (0.27-1.21) - Asian 1.03 (0.30-3.58) - Some other race 0.24 (0.08-0.76) Ordering specialty - Obstetrics/Gynecology Reference - Emergency Medicine 0.93 (0.33-2.62) - Internal Medicine 3.89 (1.48-10.20) - Family Medicine 1.62 (0.66-3.98) - Other specialty 0.87 (0.27-2.76) Has PCP 1.66 (0.81-3.42) O-RADS - 4 Reference - 5 9.15 (3.27-24.85) Imaging - MRI 7.79 (1.57-38.65) - US Reference Repeat abnormal imaging 20.61 (2.63-161.79)

Clinical risk stratification: Prague validation of the DAAE score, a clinical tool for estimating risk of disesase progression in multiple sclerosis.

Clinical risk stratification: Prague validation of the DAAE score, a clinical tool for estimating risk of disesase progression in multiple sclerosis.

Genomic profiling of colorectal liver metastasis and insight into their prognostic value.

e15536 Background: Colorectal liver metastases (CRLM) pose a significant clinical challenge, characterized by high recurrence rate even after surgical resection. Current prognostic strategies, which rely on clinicopathologic features, have demonstrated limited accuracy. As a result, there is an urgent need to identify more reliable prognostic markers. Genomic profiling of CRLM offers a promising approach to uncover molecular drivers and identify biomarkers that can improved risk stratification and guide treatment selection. Methods: In this study, we performed whole exome sequencing on 57 CRLM patients, and conducted a comprehensive analysis of primary tumor lesions. Prognostic factors associated with overall survival (OS) were systematically identified in the univariate and multivariate Cox regression analysis, and a nomogram was constructed to predict OS for CRLM patients. The nomogram further underwent external validation in an independent validation cohort of 21 CRLM patients from published data. Results: The most frequently mutated genes in our cohort were APC (64.91%) and TP53 (64.91%), followed by KRAS (50.88%), PIK3CA (24.56%) and SMAD4 (24.56%). Altered genes were primarily enriched in TP53, IGF, and Ras signaling pathway. Mutation concordance between primary and metastatic lesions was high. In the univariate analysis, a significant decrease in OS was linked with mutations in ZNF717 (HR 4.4, P = 0.00069), POTEE (HR 2.8, P = 0.015), MUC2 (HR 2.6, P = 0.017) and Amp_Chr9p13.3 (HR 1.9, P = 0.034), whereas a significant increase in OS was associated with mutations in APC (HR 0.5, P = 0.028). Multivariable analysis identified five independent prognostic factors for OS: the number of metastasis-positive lymph node stations in primary surgically removed tumor tissue, and the status of ZNF717, MUC2, APC, and Chr9p13.3 amplification. The nomogram incorporating these five factors achieved a C-index of 0.789 in this cohort, and C-index of 0.643 on an external validation cohort. Conclusions: Integration genomic profiling into routine clinical practice enables innovative prognostic strategies and improved treatment stratification for CRLM patients.

AI-based complete blood count model for colorectal cancer detection.

e15693 Background: Early colorectal cancer (CRC) detection is crucial for effective treatment; however, traditional screening methods face challenges. Colonoscopy, though effective, has limited availability, especially in resource-constrained settings. Conversely, fecal occult blood tests are cost-effective and widely available but suffer from low adherence. These barriers underscore the need for innovative approaches to support CRC screening efforts. Our retrospective study proposes developing a machine learning model based on complete blood count (CBC) tests as a risk stratification tool for CRC. By identifying high-risk individuals, this method could facilitate resource prioritization and active case-finding in at-risk populations, complementing existing screening programs and improving overall accessibility. Methods: We analyzed CBC tests from 7,588 individuals (3,990 females, 52,58%; 3,598 males, 47,42%) aged 45-75 who underwent colonoscopy or biopsies within six months of their CBC test. Among them, 468 (6,17%) were identified as cases confirmed by anatomopathological tests. The remaining 7,120 (93,83%) individuals were classified as controls, with typical colonoscopy results and without polyps, neoplasms, or other abnormalities. The database was divided into training (80%) and validation (20%) sets. The model was developed using ridge regression. Results: Descriptive analysis of all CBC biomarkers, CBC-derived ratios, and age revealed significant differences between controls and CRC cases across all features ( P < 0.001), except for lymphocytes. Using a feature selection methodology, the markers RDW, leukocytes, hemoglobin, and age were incorporated in a ridge regression model, achieving an AUC of 0.75 (95% CI: 0.74–0.75). The model was trained on a combined dataset to assess overall performance without introducing sex-based biases. It was then tested separately on male and female subsets, yielding an AUC of 0.76 (95% CI: 0.75–0.78) for males and 0.75 (95% CI: 0.74–0.75) for females, indicating consistent performance across genders. Explainability analysis revealed that increased age, higher RDW, elevated leukocyte counts, and lower hemoglobin levels were linked to a higher risk of CRC. Due to the imbalance in our dataset, we selected the threshold that maximized the balanced accuracy, resulting in a sensitivity of 0.72, specificity of 0.66, accuracy of 0.66, balanced accuracy of 0.69, negative predictive value (NPV) of 0.97, and positive predictive value (PPV) of 0.13. Conclusions: Our findings suggest that our model can contribute to the early detection of CRC, potentially serving as a valuable risk stratification tool for screening. However, the absence of external validation is a limitation of this study, requiring further research to confirm its generalizability and clinical impact.

Plasma trimethylamine levels predict adverse cardiovascular events in sudden cardiac arrest Survivors: A prospective cohort study.

Plasma trimethylamine levels predict adverse cardiovascular events in sudden cardiac arrest Survivors: A prospective cohort study.

Identification of patients at high risk for relapse by Merlin assay (CP-GEP) in an independent cohort of melanoma patients (pts) that did not undergo sentinel lymph node biopsy: An H&N subgroup analysis.

9567 Background: Sentinel lymph node biopsy (SLNB) is still the gold standard for nodal assessment used in the clinical staging of cutaneous melanoma (CM) pts by AJCC v8. Recently, we showed in a small cohort that CP-GEP also has the potential to risk stratify pts who did not undergo SLNB in low and high-risk for recurrence (Amaral et al, EJC 2023). SLNB may be challenging in pts with head and neck (H&N) melanoma, due to the regional course of cranial nerves and lymphatic drainage. Here we focus on the ability of CP-GEP to stratify pts with H&N melanoma in particular those with lentigo maligna, who did not undergo SLNB, for their risk of recurrence. Methods: We analyzed formalin-fixed paraffin-embedded primary tumor samples of 930 pts of which 206 were localized in the H&N region, with stage I/II CM diagnosed between 2000-2017 who did not receive SLNB. The CP-GEP model used combines the expression of 8 genes (SERPINE2, GDF15, ITGB3, CXCL8, LOXL4, TGFBR1, PLAT and MLANA) by quantitative reverse transcription polymerase chain reaction with age and Breslow thickness to obtain a binary output: CP-GEP Low- or High-Risk. Relapse-free survival (RFS), distant metastasis free survival (DMFS) and Melanoma Specific Survival (MSS) were evaluated using Kaplan-Meier curves. Results: We included 930 pts (stage IA-IIC) of which 206 pts (22.3%) were diagnosed with H&N melanoma. Patient characteristics: 41% were females, median age was 73-year-old, median Breslow thickness was 0.5 mm and 75.6% were lentigo maligna melanomas. Median follow up was 51 months (RFS). All H&N pts showed the following survival: 5-year RFS 82.5%, DMFS 94.0 and MSS 95.5%. CP-GEP risk stratification identified 17 patients as CP-GEP High-Risk and 188 as CP-GEP Low-Risk. The 5-year RFS rate was 86.7% for CP-GEP Low-Risk and 39.7%% for CP-GEP High-Risk pts (HR 7.85; p<0.001), 5-year DMFS was 96.3% for CP-GEP Low-Risk and 68.9% High-Risk pts (HR 10.26; p<0.001) and the 5-year MSS was 98.5% for CP-GEP Low-Risk pts and 64.7% for CP-GEP High-Risk pts (HR 24.45; p<0.01). Conclusions: Pts with H&N CP-GEP Low-Risk tumors have a good long-term survival compared to High-Risk pts even though SLN status was not assessed. This prognostic information may allow the clinicians to skip SLNB in this difficult anatomic localization and in frail and/or older pts.

Outcomes of therapy-related AML (T-AML) with venetoclax-based therapies.

6549 Background: T-AML refers to AML in patients (pts) with prior exposure to cytotoxic chemotherapy (CT) and/or radiotherapy (RT) and is often associated with adverse risk (AR) genomics. Evaluation of outcomes of T-AML with respect to type of prior therapy exposure, AML genomics, and contemporary AML therapy, especially with venetoclax (VEN), is warranted. Methods: We retrospectively analyzed pts aged ≥18 years with newly diagnosed T-AML. Pts with an antecedent myeloid disorder (MDS/CMML) prior to AML diagnosis were excluded; thus, including only pure T-AML. Composite complete response (CRc) included CR and CRi and overall response (OR) included CRc + morphologic leukemia free state. Results: From 1/2012 to 12/2023, 317 pts were included; median (med) age was 69 years (range 21-92). Overall, 120 (38%) received prior CT alone, 77 (24%) received prior RT alone (RT), and 114 (36%) received both (CRT). The most common prior malignancy was non-Hodgkin lymphoma (37%) in the CT group, prostate cancer (60%) in the RT group, and breast cancer (45%) in the CRT group.Among 286 pts with complete cytogenetic data, 180 (63%) were adverse, of whom 132 (46%) had complex karyotype (CK; 42% of CT, 48% of RT, and 61% of CRT groups). TP53 was mutated in 113/286 patients (40%) tested (36% of CT, 35% of RT, and 47% of CRT groups). Stratified by type of CT received, CK and TP53 mutation were seen in 5/5 (100%) and 3/5 (60%) of PARP inhibitor-exposed, 98/184 (53%) and 78/183 (43%) of alkylator-exposed, and 21/36 (58%) and 16/37 (43%) of topoisomerase inhibitor-exposed. Overall, 217/304 (71%) were ELN 2017 AR. In total, 251 pts (79%) received low-intensity AML therapy (LIT). CRc and OR was achieved in 122 (49%) and 146 (58%) pts treated with LIT (vs 58% and 65% with LIT+VEN). In pts treated with intensive chemotherapy (IC), the CRc and OR rate was 64% and 68% (vs 68% and 73% with IC+VEN). Overall, med RFS was 7.2 months (mos; 95% CI 5.6-8.9), and med OS was 11.8 mos (10.0-13.7). Med OS was 5.7 vs 9.0 mos (p=0.02) with LIT and LIT+VEN, respectively (resp), and med OS was 10.9 vs 48.9 mos (p=0.03) for IC vs IC+VEN, resp. Among pts treated with LIT+VEN, med OS was 14.0, 12.4, and 9.6 mos in those who had received prior CT, RT and CRT, resp; when stratified by ELN 2017 criteria, med OS was 24.6, 9.4 and 4.8 mos in the favorable, intermediate and AR groups, resp. Sixty-seven (21%) pts underwent HSCT with a landmarked comparison showing improved OS with HSCT (28.5 months vs 9.4, p<0.001). On multivariate Cox analysis in the LIT+VEN group, with forward model selection, using variables age </≥ 60, adverse cytogenetics, ASXL1, IDH1/2, FLT3-ITD, RAS, RUNX1, TP53 status , HSCT, and prior therapy group, HSCT was favorable (HR=0.19, 95% CI 0.01-0.37), along with IDH2 and NPM1 mut, while RAS and TP53 mut was associated with higher hazards of death. Other factors were not significant. Conclusions: Venetoclax improves outcomes in T-AML. In LIT+VEN treated patients, ELN 2024 risk stratification is prognostic.

Early on-treatment (on-Rx) tumor volume reduction (TVR) to predict response to the KEYNOTE-522 (KN-522) regimen in early stage triple negative breast cancer (TNBC).

592 Background: Monitoring clinical response by breast ultrasound (US) during neoadjuvant therapy is considered standard of care. We previously demonstrated that suboptimal on-Rx TVR after neoadjuvant doxorubicin and cyclophosphamide (AC) predicts non-pCR after sequential taxane-based chemo. However, it is unknown if on-Rx TVR has the similar predictive value in pts receiving the KN-522 chemo-immunotherapy regimen. Methods: Pts with early stage TNBC planned to receive the KN-522 regimen were enrolled on the prospective ARTEMIS trial (NCT02276443). Breast US was performed at baseline and after 6 weeks of paclitaxel + carboplatin + pembrolizumab. TVR was defined as the percent reduction of tumor volumes calculated using 3 perpendicular measurements of the index breast lesion. Pathological complete response (pCR) was defined as ypT0/isN0. Logistic regression was used to examine associations between covariates and pCR. Receiver operating characteristic (ROC) analyses were utilized to assess the predictive value of TVR and determine an optimal TVR threshold. Results: 150 pts were included. Clinicopathological characteristics are described in Table 1. The pCR rate was 63%. In uni- and multi-variable analyses, TVR was the only covariate to demonstrate statistically significant association with pCR (aOR:1.9 per 10% TVR, p<0.001). In ROC analyses, the area under the ROC curve (AUC-ROC) was 0.74 (95% CI: 0.66-0.82). TVR>50%, selected based on the Youden index, predicted pCR with the following performance characteristics: positive predictive value: 73%; negative predictive value: 79%; sensitivity: 94%; specificity: 41%. Conclusions: Early on-Rx TVR by breast US outperforms clinicopathological covariates in the prediction of pCR in pts with TNBC receiving the KN-522 regimen and should be leveraged for risk stratification and design of response-adapted neoadjuvant clinical trials for pts with TNBC. Clinical trial information: NCT022766443 . pCR (n=95) Non-pCR (n=55) Odds ratio (OR) p value (univariable) Adjusted OR (aOR) p value (multivariable) Median 6w TVR – % (interquartile range [IQR]) 84 (72-90) 69 (38-83) 1.5 <0.001 1.9 <0.001 Median age – years (IQR) 51 (40-61) 51 (43-64) 0.98 0.25 1.0 0.73 N (%) Ethnicity White 50 (53) 33 (60) 1 1 Black 15 (6) 10 (18) 1.1 0.84 0.9 0.91 Hispanic/Latino 25 (26) 6 (11) 2.4 0.08 1.9 0.32 Asian 5 (5) 6 (11) 0.6 0.36 0.2 0.12 T stage T1/2 79 (83) 46 (84) 1 1 T3/4 16 (17) 9 (16) 1.0 0.94 1.2 0.77 Nodal status Positive 31 (33) 24 (44) 1 1 Negative 64 (67) 31 (56) 1.60 0.18 1.4 0.55 Germline BRCA status Mutant 8 (8) 2 (3) 1 1 Wild Type 84 (88) 51 (93) 0.41 0.27 0.3 0.31 Unknown 3 (3) 2 (4) Histology Ductal 87 (92) 48 (87) 1 1 Metaplastic 3 (3) 5 (9) 0.33 0.14 0.3 0.26 Other 4 (4) 2 (4) 1.1 0.91 2.0 0.59 Unknown 1 (1) 0 Histologic grade 2 13 (14) 14 (25) 1 1 3 81 (85) 41 (75) 2.1 0.08 2.0 0.59 Unknown 1 (1) 0 Ki67 ≤35% 5 (5) 8 (15) 1 1 >35% 67 (71) 38 (69) 2.82 0.09 4.6 0.09 Unknown 23 (24) 9 (16)

Using deep learning for thyroid nodule risk stratification from ultrasound images

Using deep learning for thyroid nodule risk stratification from ultrasound images

Exploring the value of multiparametric quantitative MRI in the assessment of pancreatic ductal adenocarcinoma fibrosis grading.

To analyze the performance of multiparametric magnetic resonance imaging (MRI) in quantification of pancreatic ductal adenocarcinoma (PDAC) fibrosis grading. This prospective study enrolled 79 patients with PDAC confirmed by pathology. Multiparametric MRI including native T1 mapping, intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI), diffusion kurtosis imaging diffusion-weighted imaging (DKI-DWI), and enhanced T1 mapping were performed before surgery. Masson staining was used to evaluate intratumoral fibrosis content and classified into low- and high-fibrosis groups. MRI parameters were compared between the two groups using multivariable logistic regression analysis. The correlations between fibrosis content and MRI parameters were evaluated using Pearson's correlation. D, f, mean diffusion (MD), and enhanced T1 mapping were lower in the high-fibrosis group than in the low-fibrosis group (p < 0.001, p < 0.001, p < 0.001, p = 0.026, respectively). Native T1 mapping and extracellular volume (ECV) were opposite (All p < 0.001). No significant differences in the rest. Multivariable logistic regression revealed that native T1 mapping, MD, and ECV were independent discriminators for PDAC fibrosis grading (p = 0.037, p = 0.031, p = 0.014, respectively); the area under the curve (AUC) of native T1 mapping, MD and ECV was 0.863, 0.798, and 0.929. Among them, ECV had an extremely strong positive correlation with intratumoral fibrosis content. Native T1 mapping and MD were correlated strongly with fibrosis content (positive and negative, respectively). ECV had the highest assessing performance for grading fibrosis in PDAC compared to other MRI parameters, and has the potential to be an imaging biomarker for predicting the fibrosis content of PDAC. Question The relationship between fibrosis grade of PDAC and quantitative MRI parameters based on T1 mapping and diffusion imaging has not been fully investigated. Findings ECV performed the best in distinguishing between fibrosis grade and increased as interstitial fibrosis increased; clinical indicators offered no added value. Clinical relevance Quantitative MRI parameters provide significant value in evaluating the fibrosis grade of PDAC, which bears significant implications for preoperative risk stratification and the selection of personalized treatment strategies for patients.

Rule-in and rule-out of pre-eclampsia using a novel point-of-care placental growth factor test.

Rule-in and rule-out of pre-eclampsia using a novel point-of-care placental growth factor test.