Background/Objectives: Cabozantinib is widely used as subsequent-line therapy after immune checkpoint inhibitor (ICI) treatment in metastatic renal cell carcinoma (mRCC), yet reliable on-treatment biomarkers are lacking. This study explored the prognostic value of a composite score combining early changes in serum albumin (ΔAlb) and the systemic immune–inflammation index (ΔSII) during cabozantinib therapy. Methods: We retrospectively analyzed 40 patients with mRCC who received cabozantinib after prior ICI therapy. Alb and SII were measured at baseline and 6 weeks after initiation. Patients were stratified into three categories according to the ΔAlb + ΔSII composite: both favorable, either unfavorable, or both unfavorable. Progression-free survival (PFS) was analyzed using Kaplan–Meier and Cox regression models. Results: Among 38 evaluable patients, PFS significantly differed across composite categories (pfor trend < 0.05). Patients with both favorable changes achieved notably longer PFS, while those with both unfavorable changes experienced the shortest. Compared with the both-favorable group, the “either” and “both unfavorable” groups had shorter PFS (HR = 1.83, 95% CI 0.61–5.46; HR = 6.27, 95% CI 1.61–24.49). Conclusions: In this small retrospective cohort, early on-treatment changes in Alb and SII showed an association with PFS in ICI-pretreated mRCC treated with cabozantinib. The ΔAlb + ΔSII composite may serve as a hypothesis-generating framework, warranting confirmation in larger, prospective studies.
背景/目的:卡博替尼广泛用于转移性肾细胞癌(mRCC)免疫检查点抑制剂(ICI)治疗后的后续治疗,但目前缺乏可靠的治疗中生物标志物。本研究探讨了卡博替尼治疗期间血清白蛋白早期变化(ΔAlb)与全身免疫炎症指数早期变化(ΔSII)联合评分对预后的预测价值。方法:我们回顾性分析了40例既往接受过ICI治疗后使用卡博替尼治疗的mRCC患者。在基线及治疗开始后6周分别测量Alb和SII。根据ΔAlb + ΔSII联合评分将患者分为三类:两项均有利、任一项不利或两项均不利。采用Kaplan-Meier法和Cox回归模型分析无进展生存期(PFS)。结果:在38例可评估患者中,不同联合评分类别的PFS存在显著差异(趋势p值 < 0.05)。两项指标均呈有利变化的患者PFS显著更长,而两项均不利的患者PFS最短。与"两项均有利"组相比,"任一项不利"组和"两项均不利"组的PFS更短(HR = 1.83,95% CI 0.61–5.46;HR = 6.27,95% CI 1.61–24.49)。结论:在这项小型回顾性队列中,Alb和SII的早期治疗变化与卡博替尼治疗的ICI经治mRCC患者的PFS存在关联。ΔAlb + ΔSII联合评分可作为一个假设生成框架,值得在更大规模的前瞻性研究中加以验证。