MétaCan
Menu
Back to cohort

Construction of a predictive model for radiation proctitis after radiotherapy for female pelvic tumors based on machine learning.

2022· article· en· W4295430429 on OpenAlexaff
Hui Xie, Ming Gong, Jianfang Zhang, Qing Li

Bibliographic record

VenuePubMed · 2022
Typearticle
Languageen
FieldMedicine
TopicEffects of Radiation Exposure
Canadian institutionsArtificial Intelligence in Medicine (Canada)
Fundersnot available
KeywordsProctitisRadiation therapyMedicineLogistic regressionRadiation proctitisRandom forestMachine learningGradient boostingSupport vector machinePelvisArtificial intelligenceRadiologyComputer scienceInternal medicineDisease

Abstract

fetched live from OpenAlex

OBJECTIVES: Radiation therapy is a main method for female pelvic malignancies, which can cause some adverse reactions, such as radiation proctitis (RP). The incidence of RP is highly positively correlated with radiation dose. There is an urgent need for a scientific method to accurately predict the occurrence of RP to help doctors make clinical decisions. In this study, based on the clinical data of female pelvic tumor patients and dosimetric parameters of radiotherapy, the random forest method was used to screen the hub features related to the occurrence of RP, and then a machine learning algorithm was used to construct a risk prediction model for the occurrence of RP, in order to provide technical support and theoretical basis for the prediction and prevention of RP. METHODS: A total of 100 female patients with pelvic tumors, who received static three-dimensional conformal intensity-modulated radiation therapy in the Department of Radiation Oncology of the Affiliated Hospital of Xiangnan University from January 2019 to December 2020, were retrospectively collected, and their clinically relevant data and radiotherapy planning system data were collected. During radiotherapy and 18 months after radiotherapy, 35 cases developed RP (RP group), and the remaining 65 cases had no RP (non-RP group). The clinical and dosimetric characteristics of patients were ranked by the importance of random forest algorithm, and the independent prognostic characteristics associated with the occurrence of RP were selected for machine learning modeling. A total of 6 machine learning algorithms including support vector machines, random forests, logistic regression, lightweight gradient boosting machines, Gaussian naïve Bayes, and adaptive enhancement were used to build models. The performance of the model was evaluated by the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and F1 score. Finally, the random forest model was determined as the prediction model, and the calibration curve and decision curve of the prediction model were drawn to evaluate the accuracy and clinical benefit of the model. RESULTS: The parameters for random forest prediction model in the training set were as follow: AUC, 1.000, accuracy, 0.988, sensitivity, 1.000, specificity, 1.000, positive predictive value, 1.000, negative predictive value, 0.981, and F1 score, 1.000. In validation set, AUC was 0.713, accuracy was 0.640, sensitivity was 0.618, specificity was 0.822, positive predictive value was 0.500, negative predictive value was 0.656, and F1 score was 0.440. Random forest showed high predictive performance. Moreover, the Brief of the calibration curve for the prediction model was 0.178, the prediction accuracy was high, and the decision curve showed that the prediction model could benefit clinically. CONCLUSIONS: Based on the clinical and dosimetric parameters for the female pelvic tumor patients, the prediction model of radiation proctitis constructed by random forest algorithm has high predictive ability and strong clinical usability.

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 imitation

Not 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.

metaresearch head score (Codex)0.000
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: Simulation or modeling
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.413
Threshold uncertainty score0.513

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0000.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.000
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0000.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

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.

Opus teacher head0.010
GPT teacher head0.229
Teacher spread0.220 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one teacher head, not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designSimulation or modeling
Domainnot available
GenreEmpirical

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".

Quick stats

Citations3
Published2022
Admission routes1
Has abstractyes

Explore more

Same venuePubMedSame topicEffects of Radiation ExposureFrench-language works237,207