Urso Luca, Manco Luigi, Cittanti Corrado, Adamantiadis Sara, Szilagyi Klarisa Elena, Scribano Giovanni, Mindicini Noemi, Carnevale Aldo, Bartolomei Mirco, Giganti Melchiore
Department of Translational Medicine, University of Ferrara, Ferrara, Italy.
Nuclear Medicine Unit, Onco-Hematology Department, University Hospital of Ferrara, Via Aldo Moro 8, 44124, Ferarra, Italy.
Radiol Med. 2025 Apr;130(4):543-554. doi: 10.1007/s11547-025-01958-4. Epub 2025 Jan 28.
Build machine learning (ML) models able to predict pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) in breast cancer (BC) patients based on conventional and radiomic signatures extracted from baseline [F]FDG PET/CT.
Primary tumor and the most significant lymph node metastasis were manually segmented in baseline [F]FDG PET/CT of 52 newly diagnosed BC patients. Clinical parameters, NAC and conventional semiquantitative PET parameters were collected. The standard of reference considered was surgical pCR after NAC (ypT0;ypN0). Eight-hundred-fifty-four radiomic features (RFts) were extracted from both PET and CT datasets, according to IBSI; robust RFTs were selected. The cohort was split in training (70%) and validation (30%) sets. Four ML Models (Clinical Model, CT Model, PET Model_T and PET Model_T + N) each one with 3 learners (Random Forest (RF), Neural Network and Stochastic Gradient Descendent) were trained and tested using RFts and clinical signatures. PET Models were built considering robust RFTs extracted from either primary tumor alone (PET Model_T) or also including the reference lymph node (PET Model_T + N).
72 pathological uptakes (52 primary BC and 20 lymph node metastasis) at [F]FDG PET/CT were segmented. pCR occurred in 44.2% cases. Twelve, 46 and 141 robust RFts were selected from CT Model, PET Model_T and PET Model_T + N, respectively. PET Models showed better performance than CT and Clinical Models. The best performances were obtained by the RF algorithm of the PET Model_T + N (AUC = 0.83;CA = 0.74;TP = 78%;TN = 72%).
ML models trained on PET/CT radiomic features extracted from primary BC and lymph node metastasis could concur in the prediction of pCR after NAC and improve BC management.
基于从基线[F]FDG PET/CT中提取的传统特征和放射组学特征，构建能够预测乳腺癌（BC）患者新辅助化疗（NAC）后病理完全缓解（pCR）的机器学习（ML）模型。
在52例新诊断BC患者的基线[F]FDG PET/CT中手动分割原发性肿瘤和最显著的淋巴结转移灶。收集临床参数、NAC和传统半定量PET参数。所考虑的参考标准为NAC后的手术pCR（ypT0；ypN0）。根据国际生物医学图像标准（IBSI）从PET和CT数据集中提取854个放射组学特征（RFts）；选择稳健的RFts。将队列分为训练集（70%）和验证集（30%）。使用RFts和临床特征对四个ML模型（临床模型、CT模型、PET模型_T和PET模型_T + N）进行训练和测试，每个模型有3个学习器（随机森林（RF）、神经网络和随机梯度下降）。PET模型的构建考虑了仅从原发性肿瘤中提取的稳健RFts（PET模型_T）或还包括参考淋巴结的情况（PET模型_T + N）。
在[F]FDG PET/CT上分割出72个病理摄取（52个原发性BC和20个淋巴结转移）。44.2%的病例出现pCR。分别从CT模型、PET模型_T和PET模型_T + N中选择了12个、46个和141个稳健的RFts。PET模型表现出比CT模型和临床模型更好的性能。PET模型_T + N的RF算法获得了最佳性能（AUC = 0.83；CA = 0.74；TP = 78%；TN = 72%）。
基于从原发性BC和淋巴结转移中提取的PET/CT放射组学特征训练的ML模型可有助于预测NAC后的pCR并改善BC的管理。
