Wang Ziyang, Yuan Xiaoqiu, Nie Yuntao, Wang Jun, Jiang Guanchao, Chen Kezhong
Department of Thoracic Surgery, Peking University People's Hospital, Beijing, China.
Thoracic Oncology Institute, Peking University People's Hospital, Beijing, China.
Thorac Cancer. 2025 Mar;16(6):e70039. doi: 10.1111/1759-7714.70039.
Accurately distinguishing between multiple primary lung cancers (MPLC) and intrapulmonary metastasis (IPM) is crucial for tailoring treatment strategies and improving patient outcomes. While molecular methods offer significant advantages over traditional clinical-pathological evaluations, they lack standardized diagnostic protocols and validated prognostic value. This study systematically compared the diagnostic and prognostic performance of molecular methods versus clinical-pathological evaluations in diagnosing multiple lung cancers (MLCs), specifically focusing on the impact of next-generation sequencing (NGS) parameters on diagnostic accuracy. A review of 41 studies encompassing 1266 patients revealed that two molecular methods, Mole1 (manually counting shared mutations) and Mole2 (bioinformatics-assisted clonal probability calculation), both demonstrated superior diagnostic accuracy and prognostic discrimination capabilities. Molecular assessment, particularly Mole1, effectively stratified prognosis for MPLC and IPM, leading to significantly improved disease-free survival (DFS: HR = 0.24, 95% CI: 0.15-0.39) and overall survival (OS: HR = 0.33, 95% CI: 0.18-0.58). Further analysis suggests that a minimal panel of 30-50 genes may be sufficient to effectively differentiate prognoses. Compared to Mole1, Mole2 demonstrated greater specificity and stability across various panels, achieving AUC values from 0.962 to 0.979. Clinical-pathological evaluations proved unreliable, not only failing to distinguish prognosis effectively but also exhibiting a potential misdiagnosis rate of 35.5% and 33.6% compared to the reference diagnosis. To improve both cost-effectiveness and diagnostic accuracy, bioinformatics-assisted molecular diagnostics should be integrated into multidisciplinary assessments, especially for high-risk cases where diagnostic errors are common.
准确区分多原发性肺癌（MPLC）和肺内转移（IPM）对于制定治疗策略和改善患者预后至关重要。虽然分子方法相较于传统临床病理评估具有显著优势，但它们缺乏标准化的诊断方案和经过验证的预后价值。本研究系统地比较了分子方法与临床病理评估在诊断多原发性肺癌（MLC）中的诊断和预后性能，特别关注下一代测序（NGS）参数对诊断准确性的影响。对41项研究（涉及1266例患者）的回顾显示，两种分子方法，即Mole1（手动计数共享突变）和Mole2（生物信息学辅助克隆概率计算），均表现出卓越的诊断准确性和预后判别能力。分子评估，尤其是Mole1，有效地对MPLC和IPM的预后进行了分层，使无病生存期（DFS：HR = 0.24，95% CI：0.15 - 0.39）和总生存期（OS：HR = 0.33，95% CI：0.18 - 0.58）得到显著改善。进一步分析表明，一个包含30 - 50个基因的最小基因 panel 可能足以有效区分预后。与Mole1相比，Mole2在不同基因 panel 中表现出更高的特异性和稳定性，AUC值在0.962至0.979之间。临床病理评估被证明不可靠，不仅无法有效区分预后，而且与参考诊断相比，潜在误诊率分别为35.5%和33.6%。为了提高成本效益和诊断准确性，应将生物信息学辅助分子诊断纳入多学科评估，特别是对于诊断错误常见的高危病例。
