Abstract CT088: Clinical validation of a novel blood-based protein multi-analyte test for early detection of pancreatic ductal adenocarcinoma (PDAC) in a large, independent high-risk patient population
Bibliographic record
Abstract
Abstract BACKGROUND: Pancreatic cancer is the third highest cause of cancer mortality in the United States. Detecting PDAC at an earlier stage with the tumor confined to the pancreas and lymph node negative improves 5-year rates from 3% for late-stage diagnosis to 44%. There is no FDA approved early detection blood test for PDAC. IMMNOV-2, a blood-based protein biomarker model comprising ICAM1, TIMP1, CTSD, THBS1 and CA19-9 was previously shown to differentiate early-stage PDAC from high-risk controls with high sensitivity and specificity. The current study aimed to validate the performance of IMMNOV-2 in detecting early-stage pancreatic cancer in a large clinical population independent from which the model was developed. METHODS: This was a multi-institutional blinded study assessing the performance of IMMNOV-2 in patient serum samples to differentiate Stage I and Stage II PDAC cases from non-PDAC controls at high-risk due to familial, genetic, or clinical factors. The model’s performance in the whole patient population was also compared to CA19-9 performance alone. IMMNOV-2 comprises four quantitative ELISAs that measure the concentration of the protein biomarkers in human serum. CA19-9 is measured using a Roche COBAS. A fixed mathematical algorithm is employed to integrate the values of the five biomarkers to calculate a positive or negative call based on a predefined cutoff. RESULTS: 202 Stage I and II PDACs and 864 high-risk controls were enrolled. Assays were performed in a blinded manner. IMMNOV-2 distinguished early-stage PDAC from high-risk controls with 78.2% (95% CI, 71.9-83.7) sensitivity at 93.5% (95% CI, 91.7-95.1) specificity. In contrast, CA19-9 alone differentiated early-stage PDAC from controls with 64% (95% CI, 57.3-71) sensitivity (p<0.001 vs IMMNOV-2) at 94.7% (95% CI, 93-96.1) specificity. Performance between Stage I and Stage II PDAC cases were similar. A pre-planned analysis revealed a decrease of IMMNOV-2 performance with increasing age of samples. In samples collected <5 years before the study (89 cases, 751 high-risk controls), sensitivity and specificity of the test was 82.0% (95% CI, 74-90) and 94.9% (95% CI, 93.1-96.4), respectively, which was significantly better than CA19-9 alone (p<0.001) and performance in samples collected >5 yrs before the study (sensitivity 75.2% (95% CI, 67.3-83.2), specificity 84.0% (95% CI, 77.3-90.8), p<0.001). CONCLUSION: IMMNOV-2 differentiated Stage I and Stage II PDAC from high-risk controls with high accuracy in this large clinical validation study. Model performance was significantly better than CA19-9 alone. Testing in recently collected samples would be consistent with clinical use of the test and results in better performance for detecting early-stage PDAC. These promising data warrant further validation in a next-level study using prospective randomized open blinded endpoint (PROBE) design principles. Citation Format: Bryson Katona, Norma Alonzo Palma, Aimee Lucas, Rosalie Sears, Salvatore Paiella, George Zogopoulos, Eli M. Grindedal, Raymond Wadlow, Erkut Borazanci, Daniel A. Sussman, Ora Gordon, Natasha Kureshi, Lisa Ford, Thomas King, Randall Brand, Diane Simeone. Clinical validation of a novel blood-based protein multi-analyte test for early detection of pancreatic ductal adenocarcinoma (PDAC) in a large, independent high-risk patient population [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_2):Abstract nr CT088.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".