Abstract：
BACKGROUND: Approximately 30% of patients with clinical stage I non-seminoma (CSI-NS) relapse. Current risk stratification is based on lymphovascular invasion (LVI) alone. The extent to which additional tumor characteristics can improve risk prediction remains unclear.
OBJECTIVE: To determine the most important prognostic factors for relapse in CSI-NS patients.
METHODS: Population-based cohort study including all patients with CSI-NS diagnosed in Denmark between 2013 and 2018 with follow-up until 2022. Patients were identified in the prospective Danish Testicular Cancer database. By linkage to the Danish National Pathology Registry, histological slides from the orchiectomy specimens were retrieved.
METHODS: Histological slides were reviewed blinded to the clinical outcome. Clinical data were obtained from medical records. The association between prespecified potential prognostic factors and relapse was assessed using Cox regression analysis. Model performance was evaluated by discrimination (Harrell\'s C-index) and calibration.
RESULTS: Of 453 patients included, 139 patients (30.6%) relapsed during a median follow-up of 6.3 years. Tumor invasion into the hilar soft tissue of the testicular hilum, tumor size, LVI and embryonal carcinoma were independent predictors of relapse. The estimated 5-year risk of relapse ranged from < 5% to > 85%, depending on the number of risk factors. After internal model validation, the model had an overall concordance statistic of 0.75. Model calibration was excellent.
CONCLUSIONS: The identified prognostic factors provide a much more accurate risk stratification than current clinical practice, potentially aiding clinical decision-making.
OBJECTIVE: To determine the most important prognostic factors for relapse in CSI-NS patients.
METHODS: Population-based cohort study including all patients with CSI-NS diagnosed in Denmark between 2013 and 2018 with follow-up until 2022. Patients were identified in the prospective Danish Testicular Cancer database. By linkage to the Danish National Pathology Registry, histological slides from the orchiectomy specimens were retrieved.
METHODS: Histological slides were reviewed blinded to the clinical outcome. Clinical data were obtained from medical records. The association between prespecified potential prognostic factors and relapse was assessed using Cox regression analysis. Model performance was evaluated by discrimination (Harrell\'s C-index) and calibration.
RESULTS: Of 453 patients included, 139 patients (30.6%) relapsed during a median follow-up of 6.3 years. Tumor invasion into the hilar soft tissue of the testicular hilum, tumor size, LVI and embryonal carcinoma were independent predictors of relapse. The estimated 5-year risk of relapse ranged from < 5% to > 85%, depending on the number of risk factors. After internal model validation, the model had an overall concordance statistic of 0.75. Model calibration was excellent.
CONCLUSIONS: The identified prognostic factors provide a much more accurate risk stratification than current clinical practice, potentially aiding clinical decision-making.
摘要：
背景：大约30%的临床I期非精原细胞瘤（CSI-NS）患者复发。当前的风险分层仅基于淋巴管浸润（LVI）。其他肿瘤特征在多大程度上可以改善风险预测仍不清楚。
目的：确定CSI-NS患者复发的最重要预后因素。
方法：基于人群的队列研究，包括2013年至2018年在丹麦诊断为CSI-NS的所有患者，随访至2022年。在前瞻性丹麦睾丸癌数据库中确定了患者。通过与丹麦国家病理学登记处的联系，从睾丸切除术标本中检索组织学切片。
方法：对临床结果不了解的组织学切片进行回顾。从病历中获得临床数据。使用Cox回归分析评估预设的潜在预后因素与复发之间的关联。通过辨别(Harrell'sC指数)和校准评估模型性能。
结果：包括453例患者，139例患者（30.6%）在中位随访6.3年期间复发。肿瘤侵入睾丸门的肺门软组织，肿瘤大小,LVI和胚胎癌是复发的独立预测因子。估计的5年复发风险从<5%到>85%不等。取决于风险因素的数量。内部模型验证后，该模型的总体一致性统计量为0.75.模型校准非常好。
结论：与目前的临床实践相比，确定的预后因素提供了更准确的风险分层，可能有助于临床决策。
目的：确定CSI-NS患者复发的最重要预后因素。
方法：基于人群的队列研究，包括2013年至2018年在丹麦诊断为CSI-NS的所有患者，随访至2022年。在前瞻性丹麦睾丸癌数据库中确定了患者。通过与丹麦国家病理学登记处的联系，从睾丸切除术标本中检索组织学切片。
方法：对临床结果不了解的组织学切片进行回顾。从病历中获得临床数据。使用Cox回归分析评估预设的潜在预后因素与复发之间的关联。通过辨别(Harrell'sC指数)和校准评估模型性能。
结果：包括453例患者，139例患者（30.6%）在中位随访6.3年期间复发。肿瘤侵入睾丸门的肺门软组织，肿瘤大小,LVI和胚胎癌是复发的独立预测因子。估计的5年复发风险从<5%到>85%不等。取决于风险因素的数量。内部模型验证后，该模型的总体一致性统计量为0.75.模型校准非常好。
结论：与目前的临床实践相比，确定的预后因素提供了更准确的风险分层，可能有助于临床决策。
