Test characteristics of end-of-treatment PET scans in diffuse large b-cell lymphoma in a population-level cohort
Bibliographic record
Abstract
Abstract Introduction: First-line treatment of diffuse large B-cell lymphoma (DLBCL) consists of chemoimmunotherapy followed by end-of-treatment (EOT) positron-emission tomography (PET) scan to assess response. However, EOT PET can have false-positive results, thus, positive results must be followed by further testing to clarify response status, either by tissue biopsy or repeat PET. Guidelines suggest waiting preferably at least 6 weeks from last chemotherapy to perform PET, to avoid false positives from post-treatment inflammation. However, prior studies examining the false positive rate of EOT PET have small sample sizes and have not examined the impact of timing of the EOT PET. We conducted an exploratory analysis of EOT PET test characteristics using population-level data and evaluated the impact of time to EOT PET. Methods: We conducted a retrospective cohort study using population-level administrative health data of all patients in Ontario, Canada, ≥18 years old with DLBCL, who received frontline rituximab-based chemoimmunotherapy followed by PET within 16 weeks of last chemotherapy from October 2009-May 2021. We evaluated the time from last chemoimmunotherapy to EOT PET, with short time to PET defined as ≤6 weeks, and long time to PET >6 weeks. The presence of follow-up testing, defined as either repeat PET or tissue biopsy within 6 months of EOT PET was used as a surrogate for a positive EOT PET. True positive was defined as the presence of follow-up testing, as well as either cancer relapse or cancer-specific death within 2 years of the EOT PET. False positive was follow-up testing in the absence of relapse or cancer-specific death within 2 years. True negative was defined as no follow-up testing and no relapse or cancer-specific death within 2 years. False negative was defined as no follow-up testing, and relapse or cancer-specific death within 2 years of the EOT PET. These definitions were used to calculate the positive and negative predictive values, sensitivity and specificity of EOT PET both for the entire cohort, and according to time to EOT PET. Results: A total of 3833 individuals with DLBCL had EOT PET scan following first-line treatment; 2247 (59%) had short time to PET (≤6 weeks), 1586 (41%) had long time to PET (>6 weeks). For the overall cohort, 715 (19%) patients had a false positive EOT PET, 428 (11%) had true positive, 358 (9%) had false negative, and 2332 (61%) had true negative EOT PET. In the group with short time to EOT PET (≤6 weeks), there were 457 (20%) false positives, and 185 (8%) false negatives. In the group with long time to EOT PET (>6 weeks), there were 258 (16%) false positives, and 173 (11%) false negatives. The positive predictive value (PPV) was 37% for the overall cohort, 37% for the short time to EOT PET group, and 39% in the long time to EOT PET group. The negative predictive value (NPV) was 87% for the whole cohort, 88% in the short time to EOT PET group, and 85% in the long time to EOT PET group. Specificity was 76% for the whole cohort, 75% for short time to EOT PET, and 79% for long time to EOT PET. Sensitivity was 54% for the whole cohort, 59% for short time to EOT PET, and 49% for long time to EOT PET. Conclusions:Positive predictive value of EOT PET is suboptimal, and a substantial number of patients require follow up testing despite not ultimately relapsing. Waiting a longer time to PET (>6 weeks) results in slightly lower false positives, higher PPV and specificity, with slightly lower NPV and sensitivity.
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 | 0.001 | 0.006 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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".