PREC Multicentre Restrospective Study: A Preoperative Risk Classification for Synchronous Oligometastatic Non-Small Cell Lung Cancer

In our previous multicentre study (Lung Cancer, 2021;154:29-35), longest survival was observed in synchronous lung oligometastatic/non pN2 disease. Aims. To identify preoperative favourable prognostic factors and to propose a preoperative classification for categorising the synchronous oligometastatic NSCLC. Retrospectively review of records (2005–2018). Inclusion criteria: synchronous oligometastatic NSCLC (≤5 extrapulmonary metastases), radical surgical treatment of primary tumour radical and all metastatic sites with/without neoadjuvant/adjuvant therapy. Exclusion criteria: palliative/diagnostic surgery, relapsing lung cancer, no follow-up information. Primary endpoint: identification of risk classification for oligometastatic NSCLC. Statistical analysis. Median OS/PFS were estimated by reverse Kaplan–Meier method. Stratified backward stepwise Cox regression model was employed for multivariable survival analyses. Backward elimination was performed with p-value of 0.20. Akaike information criterion was used to estimate the models' relative quality, selecting the ones with the best goodness of fit and avoiding collinearity bias. A prognostic grouping that could consider all the relevant prognostic factors simultaneously was constructed. ROC curve was generated. Hosmer-Lemeshow χ statistics was used for measuring OS calibration within groups. 281 patients included (Table 1a). Data from the Cox regression model (Table 1b) was used to construct a prognostic risk classification.Table 1aCharacteristics of the populationVariablesNo. (%)Age (years), mean ± SD62.7 ± 9.6Male/female ratio1.8Tumour descriptor - cT1a - cT1b - cT1c - cT2a - cT2b - cT3 - cT4 Node descriptor - cN0 - cN1 - cN29 (3.3) 40 (14.5) 87 (31.5) 44 (15.9) 43 (15.6) 35 (12.7) 18 (6.5) 155 (56.2) 46 (16.7) 75 (27.2)Metastasis location - Brain - Adrenal gland(s) - Contralateral lung - Bone - Subcutaneous tissue - Other sites143 (50.9) 46 (16.4) 40 (14.2) 36 (12.8) 4 (1.4) 3 (1.1)Histological diagnoses - Adenocarcinoma - Squamous-cell carcinoma - Others - Adenosquamous carcinoma219 (77.9) 33 (11.7) 11 (3.9) 9 (3.2) Open table in a new tab Table 1bResults of Cox regression analysis for progression-free survivalVariablesHR95% CIp-valueAge >651.371.00 – 1.570.0269cN1/cN22.100.68 – 6.440.019Lung/brain metastases2.961.44 – 6.070.0031Induction treatment1.380.56 – 3.400.027Four parameters (age, site of metastasis, clinical nodal status, and induction treatment) were used to build a risk classification (Figure): • Group A. No risk factors (age ≤65 years, lung/brain metastasis, cN0, induction treatment). • Group B. One risk factor (age >65 years or no lung/brain metastasis or cN1–2 or no induction treatment). • Group C. Two risk factors. • Group D. ≥3 risk factors. AUC was 0.56 (95% CI: 0.49–0.62), Hosmer-Lemeshow χ statistics was 21.3 (3 degrees of freedom, p=0.0042). Open table in a new tab Four parameters (age, site of metastasis, clinical nodal status, and induction treatment) were used to build a risk classification (Figure): • Group A. No risk factors (age ≤65 years, lung/brain metastasis, cN0, induction treatment). • Group B. One risk factor (age >65 years or no lung/brain metastasis or cN1–2 or no induction treatment). • Group C. Two risk factors. • Group D. ≥3 risk factors. AUC was 0.56 (95% CI: 0.49–0.62), Hosmer-Lemeshow χ statistics was 21.3 (3 degrees of freedom, p=0.0042). Patient selection is critical in identifying the proper subsets of oligometastatic NSCLC. After validation, this preoperative risk classification might support decision-making during the multidisciplinary team assessments and patients' selection for enrollment in future randomised trials.