Comparative Analysis of Atlas and Neural Network Autosegmentation Methods for Pediatric Craniospinal Irradiation With the Development of a Knowledge-Based Quality Assurance Tool
Bibliographic record
Abstract
Purpose: This study aims to evaluate the performance of Atlas and neural network autosegmentation methods and develop a knowledge-based quality assurance (QA) tool for pediatric craniospinal irradiation (CSI). Methods and Materials: Autosegmentation was performed on 63 CSI patients using 3 methods: Atlas, commercial artificial intelligence (AI), and in-house AI. The performance of these methods was analyzed using 13 quantitative metrics, comprising 6 overlap and 7 distance metrics, across 13 critical organs and a linear mixed-effect model analysis was performed. Additionally, a knowledge-based QA tool was developed by leveraging distinctive computed tomography number distributions from 100 CSI patients for each organ, using the kernel density estimation (KDE) method to ensure robust error detection and validation. The QA tool was tested on 50 CSI cases by comparing baseline KDEs from 100 CSI patients. Results: The linear mixed-effect analysis showed that the in-house AI outperformed both the Atlas and commercial AI methods in overlap and distance metrics. The in-house AI outperformed the commercial AI with a higher average overlap of 0.01 ± 0.01 and surpassed the Atlas method by 0.02 ± 0.01. In terms of distance metrics, the in-house AI matched the commercial AI (-0.31 ± 0.72 mm) and exceeded the Atlas method by 3.10 ± 0.68 mm. Paired t-tests showed the in-house AI was superior to the Atlas in 13.0% of cases, while the Atlas outperformed the in-house method in 8.9% of comparisons. Similarly, the in-house AI was better than the commercial AI in 35.3% of tests, with the commercial AI outperforming in 32.7%. The QA tool results demonstrated that 100% agreement with baseline KDEs occurred in 46.4% of tests for Atlas, 46.5% for the commercial AI, and 60.7% for the in-house AI. Conclusions: The in-house AI excelled over the Atlas and commercial AI methods in autosegmentation accuracy for pediatric CSI patients. Furthermore, a knowledge-based QA tool enables clinicians to detect and correct gross errors in autosegmentation.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".