Building a model to predict the risk of multiple severe neurotoxicities in cancer survivors after cisplatin treatment.
Bibliographic record
Abstract
e24066 Background: Cisplatin treatment is used for many cancers, including testicular, ovarian, and head and neck malignancies. Cancer survivors with multiple cisplatin-related toxicities can have poor health-related quality of life (HRQOL). Identification of clinical and genetic factors that predict the risk of these neurotoxicities is critical. Methods: Testicular cancer survivors (TCS) enrolled in the Platinum Study completed surveys, underwent physical examination, extensive audiometric testing, and phlebotomy for genotyping and serum platinum analysis. Cases included TCS with two or more severe toxicities (hearing loss [HL], tinnitus, and peripheral sensory neuropathy [PSN]), defined as follows: hearing threshold > 40dB based on geometric mean of 4-12kHz, responding yes to “Do you have ringing or buzzing in the ears?” and/or EORTC-CIPN20 scores in the severe range for items related to sensory neuropathy. Controls were restricted to TCS without any toxicities. TCS with a single toxicity were excluded from analyses. Penalized logistic regression lasso method was used to create the model to predict the binary outcome. Creatinine clearance and residual serum platinum levels were calculated. Polygenic risk scores (PRS) for traits commonly associated with pharmacokinetics and HL, tinnitus, and PSN were calculated for TCS in the training (n = 284) and validation (n = 157) data sets using PRS publicly available in The Polygenic Score Catalog using PRSice 2.3.3. Models were trained and tested in R 4.1.2. Results: A model to assess the risk of developing multiple severe neurotoxicities that could be used without blood work and additional analysis was developed. Clinical predictors incorporated into the model were age at testicular cancer diagnosis, age at phlebotomy, weight and height. PRS incorporated were age-related sensorineural hearing loss (PGS000762), body fat percentage (PGS002133), creatinine in urine (PGS001944), and peripheral nervous system disease (PGS002039). The accuracy of this model was 77.71%, which was significantly greater than the no information rate (NIR) of 65.61% (p = .00067). The positive and negative predictive values (PPV and NPV) were 72.09% and 79.82%, respectively. The AUC-ROC was 0.804. Adding residual platinum levels and creatinine clearance increased the accuracy of the model to 78.34%, which was significantly greater than the NIR (p = .00035). The PPV was 75.00% and the NPV was 79.49%. The area under the receiver operating characteristic curve (AUC-ROC) was 0.832. Conclusions: TCS are often faced with multiple severe neurotoxicities such as HL, tinnitus, and PSN, which impact HRQOL for many decades. If confirmed, a penalized regression model using clinical and genetic characteristics can predict the risk of developing these phenotypes to guide clinicians in treatment and post-treatment management plans.
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 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.009 | 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".