Natural course of lung adenocarcinoma manifesting as ground-glass nodules: invasiveness assessment based on growth evaluation.

Background: Investigating the growth characteristics of lung adenocarcinoma manifesting as ground-glass nodules (gLUAD) and assessing its invasiveness based on these features are crucial for optimizing follow-up and intervention strategies. This study aimed to systematically analyze the growth dynamics of gLUAD and compare the value of different growth evaluation methods in predicting gLUAD invasiveness.

Methods: A total of 564 participants with 625 gLUAD were retrospectively enrolled from the First and Fourth Medical Centers of the Chinese PLA General Hospital between January 2018 and December 2022. gLUAD was pathologically classified into two categories: adenocarcinoma in situ (AIS)/minimally invasive adenocarcinoma (MIA) and invasive adenocarcinoma (IAC). Growth evaluation was conducted using four definitions (size growth, volume growth, mass growth, and stage shift) and two growth models (linear and exponential models). These methods were further evaluated in terms of their effectiveness in assessing gLUAD invasiveness.

Results: The median follow-up period of gLUAD in this study was 1,050 days. The median doubling times for total/solid size, volume, and mass of gLUAD were >3,650/3,042 days, 1,460/1,014 days, and 1,521/1,014 days, respectively. For AIS/MIA, all doubling times exceeded 3,650 days, whereas for IAC, they were 3,318/2,147 days, 1,141/777 days, and 1,074/760 days, respectively. The R² and root mean squared error (RMSE) values for the linear and exponential growth models in fitting total/solid size, volume, and mass were 0.98/0.95 vs. 0.98/0.95, 0.98/0.90 vs. 0.98/0.93, and 0.95/0.90 vs. 0.98/0.92 for R², while they were 0.53/0.92 vs. 0.51/0.94, 169.80/100.00 vs. 143.87/106.81, and 116.12/101.24 vs. 72.81/108.99 for RMSE, respectively. The median growth times for size growth, volume growth, mass growth, and stage shift of gLUAD were 1,273, 750, 792, and 1,672 days, respectively. Compared to AIS/MIA, IAC exhibited significantly higher growth rates (all P values <0.001). For invasiveness assessment, the linear growth rates (LGRs) outperformed the exponential growth rates (EGRs) in discriminative value (all P values <0.05). Among the growth evaluation methods, the LGR of total mass demonstrated the highest discriminative ability for pathological subtypes of gLUAD, with an area under the curve (AUC) of 0.83. At an optimal cutoff of 16.80 mg/year, the accuracy, sensitivity, and specificity for distinguishing gLUAD with different invasiveness were 0.78, 0.77, and 0.79, respectively.

Conclusions: gLUAD exhibited an indolent growth pattern, with significant differences in growth trends between AIS/MIA and IAC. Both the linear and exponential growth models showed similar fitting performance for gLUAD growth, while the linear growth model provided a more reliable assessment of invasiveness. The LGR of total mass offered superior differentiation between AIS/MIA and IAC and may serve as a valuable auxiliary indicator for assessing gLUAD invasiveness.