Citations

Accurate prediction of progression in subjects at risk of Alzheimer's disease is crucial for enrolling the right subjects in clinical trials. However, a prospective comparison of state-of-the-art algorithms for predicting disease onset and progression is currently lacking. We present the findings of "The Alzheimer's Disease Prediction Of Longitudinal Evolution" (TADPOLE) Challenge, which compared the performance of 92 algorithms from 33 international teams at predicting the future trajectory of 219 individuals at risk of Alzheimer's disease. Challenge participants were required to make a prediction, for each month of a 5-year future time period, of three key outcomes: clinical diagnosis, Alzheimer's Disease Assessment Scale Cognitive Subdomain (ADAS-Cog13), and total volume of the ventricles. The methods used by challenge participants included multivariate linear regression, machine learning methods such as support vector machines and deep neural networks, as well as disease progression models. No single submission was best at predicting all three outcomes. For clinical diagnosis and ventricle volume prediction, the best algorithms strongly outperform simple baselines in predictive ability. However, for ADAS-Cog13 no single submitted prediction method was significantly better than random guesswork. Two ensemble methods based on taking the mean and median over all predictions, obtained top scores on almost all tasks. Better than average performance at diagnosis prediction was generally associated with the additional inclusion of features from cerebrospinal fluid (CSF) samples and diffusion tensor imaging (DTI). On the other hand, better performance at ventricle volume prediction was associated with inclusion of summary statistics, such as the slope or maxima/minima of patient-specific biomarkers. On a limited, cross-sectional subset of the data emulating clinical trials, performance of the best algorithms at predicting clinical diagnosis decreased only slightly (2 percentage points) compared to the full longitudinal dataset. The submission system remains open via the website https://tadpole.grand-challenge.org, while TADPOLE SHARE (https://tadpole-share.github.io/) collates code for submissions. TADPOLE's unique results suggest that current prediction algorithms provide sufficient accuracy to exploit biomarkers related to clinical diagnosis and ventricle volume, for cohort refinement in clinical trials for Alzheimer's disease. However, results call into question the usage of cognitive test scores for patient selection and as a primary endpoint in clinical trials.

4525 Background: Pancreatic cancer cells and surrounding stroma are known to overexpress SPARC (secreted protein acid rich in cysteine), which is associated with poor clinical outcomes. nab-P, an albumin-bound nanoparticle form of paclitaxel increased tumor accumulation of paclitaxel through binding of albumin to SPARC. This disease specific phase 1 study was designed to evaluate the safety and efficacy of G + nab-P and the correlation of response with tumor SPARC and serum CA19–9 levels. Methods: nab-P doses (100–150 mg/m2) + (G) (1000 mg/m2) were given on days 1, 8, and 15 of a 28-day cycle to pts with metastatic pancreatic adenocarcinoma and with no prior chemotherapy for their metastatic disease. Level 3 SPARC staining by immunohistochemistry was considered positive. Results: 63 pts received treatment. The most common grade 3 and 4 adverse event that occurred in >20% of pts was neutropenia. Nine (18%) pts and 4 (8%) pts had a grade 3/4 event, respectively. Neuropathy was also observed. One combination-associated death due to sepsis occurred at the 150 mg/m2 nab-P level. Serial PET scans of 53 pts with outside adjudication to date showed 12 (23%) complete responses, 29 (55%) partial responses (PRs) and 4 (8%) stable disease (SD). By RECIST criteria, of the 49 pts evaluable to date, 1(2%) had CR, 12 (24%) had PR, and 20 (41%) had SD. The median survival was 9 months to date. SPARC data were available for 35 pts, of which 10 (29%) were SPARC+ and 25 (71%) were SPARC-. Of these, 27 pts had evaluable response data. Pts that were SPARC+ (8/27) were more likely to be responders (6/8, 75%) than pts who were SPARC- (5/19, 26%), P = 0.03, Fisher's exact test. Median progression-free survival (PFS) increased from 4.8 months for SPARC- pts (22 pts) to 6.2 months for SPARC+ pts (9 pts); however, these data are still immature. Of 45 pts with elevated CA19–9 at baseline, 42 (93%) had maximum decrease of >40% with a median maximum decrease of 92%. Conclusions: The combination of nab-P and G was generally well tolerated and had substantial enough antitumor activity in patients with pancreatic cancer to warrant a phase III clinical trial. SPARC+ status in these patients was associated with higher response rate and longer PFS. [Table: see text]