Abstract PD5-05: Including a 21-gene assay recurrence score in multivariable predictive model generation improves prediction of local recurrence after breast conserving surgery for ductal carcinoma-in-situ
Notice bibliographique
Résumé
Abstract Introduction Accurate prediction of local recurrence (LR) after breast conserving surgery (BCS) for ductal carcinoma-in-situ (DCIS) is crucial to personalize recommendations for adjuvant radiotherapy (RT). The 21-gene recurrence score (RS) predicts for distant metastases for woman with invasive breast cancer. We hypothesised that the RS could improve prediction of LR after BCS for DCIS. Methods We performed a population-based analysis of 1226 woman aged ≤74 treated with BCS ± RT for pure DCIS. Expert pathology review was obtained for all cases, as was the RS. Treatment and outcomes were obtained by deterministic linkage to administrative databases and chart review. Clinico-pathologic features obtained included: age, tumor size, nuclear grade, presence of comedonecrosis, multifocality, margins, and adjuvant radiation. The outcome assessed was local recurrence by 10 years from diagnosis of DCIS. The LR prediction model was developed using multivariable Cox regression, where a non-parametric approach was implemented to estimate the baseline hazard function. The proportional hazards assumption was assessed and time-interaction terms were included with each covariate in the model. Models were ranked based on c-statistic, log-likelihood estimate, and Akaike information criterion (AIC). Backward selection was used to obtain the final reduced model with time-interaction terms. Calibration for the best model was examined by grouping predicted 10-year risk of LR into deciles and plotting against observed 10-year risk of LR based on mean Kaplan-Meier estimates. Internal validation was performed by bootstrapping. Results Of the 1226 woman included, 514 were treated with BCS alone and 712 received adjuvant RT. Median follow up from time of treatment was 16 years (interquartile range (IQR): 14-18). The median age was 56 years (IQR: 49-64). Margins were negative in 90.5% of cases. Tumor size was ≤1cm in 430 (35.1%), 1-2.5cm in 633 (51.6%), and >2.5cm in 163 (13.3%). The median RS was 15 (IQR: 8-30) and the mean RS was 21.37 (SD 18.93). The best predictive model included the RS and had a c-statistic of 0.68 as well as the lowest AIC. This model included the following variables: RS, age, tumor size, nuclear grade, margin status, comedonecrosis (≤30% vs higher), multifocality, and treatment (BCS vs BCS+RT); it also included the following interaction terms: treatment and time, RS and time, comedonecrosis and time, and treatment and tumor size. Due to the non-linear relationship between certain characteristics and the risk of LR, quadratic terms for RS and age were also included. This model was well calibrated overall, especially in the lower risk range around the 10% risk threshold. It was also well calibrated in this risk range in the subset of woman who were treated with BCS alone. Conclusion The best performing model generated to predict LR after BCS for DCIS includes the RS. Work is ongoing to compare RS and the 12-gene DCIS score in terms of prediction of LR, as well as prediction of invasive LR specifically. This work can help guide future clinical de-escalation trials by better identifying woman with truly low risk of LR after BCS for DCIS. Citation Format: Ezra Hahn, Rinku Sutradhar, Sumei Gu, Lawrence Paszat, Danielle Rodin, Sharon Nofech-Mozes, Cindy Fong, Eileen Rakovitch. Including a 21-gene assay recurrence score in multivariable predictive model generation improves prediction of local recurrence after breast conserving surgery for ductal carcinoma-in-situ [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PD5-05.
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction machine sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Le volet Gemma est une étiquette directe du modèle pour chaque travail de la base, lue sur la notice réduite au titre. Le volet Codex est un classifieur appris des 10 348 étiquettes directes de Codex et calibré sur les taux pondérés de l'échantillon; les champs sans appui suffisant ne portent aucun appel Codex. Le mode candidate est l'union des deux volets; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont pas des étiquettes humaines.
Scores du classifieur distillé par catégorie (deux têtes)
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,009 | 0,020 |
| Méta-épidémiologie (sens strict) | 0,001 | 0,000 |
| Méta-épidémiologie (sens large) | 0,001 | 0,001 |
| Bibliométrie | 0,001 | 0,001 |
| Études des sciences et des technologies | 0,000 | 0,000 |
| Communication savante | 0,001 | 0,000 |
| Science ouverte | 0,001 | 0,001 |
| Intégrité de la recherche | 0,000 | 0,001 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,004 | 0,000 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule source (Gemma direct ou Codex distillé), pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».