Distinctive pathological and clinical features of lung carcinoids with high proliferation index

Abstract

Typical (TCs) and atypical carcinoids (ACs) are defined based on morphological criteria, and no grading system is currently accepted to further stratify these entities. The 2015 WHO classification restricts the Ki-67 role to biopsy or cytology samples, rather than for prognostic prediction. We aimed to investigate whether values and patterns of Ki-67 alone would allow for a clinically meaningful stratification of lung carcinoids, regardless of histological typing. Ki-67 proliferation index and pattern (homogeneous versus heterogeneous expression) were assessed in a cohort of 171 TCs and 68 ACs. Cases were subdivided into three Ki-67 ranges (<4/4-9/≥10%). Correlations with clinicopathological data, univariate and multivariate survival analyses were performed. The majority of cases (61.5%) belonged to the <4% Ki-67 range; 25.1 and 13.4% had a proliferation index of 4-9% and ≥10%, respectively. The <4% Ki-67 subgroup was significantly enriched for TCs (83%, p < 0.0001); ACs were more frequent in the subgroup showing Ki-67 ≥ 10% (75%, p < 0.0001). A heterogeneous Ki-67 pattern was preferentially seen in carcinoids with a Ki-67 ≥10% (38%, p < 0.02). Mean Ki-67 values ≥4 and ≥10% identified categories of poor prognosis both in terms of disease-free and overall survival (p = 0.003 and <0.0001). At multivariate analysis, the two thresholds did not retain statistical significance; however, a Ki-67 ≥ 10% identified a subgroup of dismal prognosis even within ACs (p = 0.03) at univariate analysis. Here, we describe a subgroup of lung carcinoids showing brisk proliferation activity within the necrosis and/or mitotic count-based categories. These patients were associated with specific clinicopathological characteristics, to some extent regardless of histological subtyping.

Keywords: Carcinoid; Heterogeneity; Ki-67; Lung; Prognosis; WHO classification.

Fig. 1

Atypical carcinoids showing different Ki-67 patterns (homogenous versus heterogeneous). ACs with organoid and trabecular growth patterns (a and c: H&E 40×, inset 100×) showing a homogenously (b: 40×) and heterogeneously (d: 40×) increased Ki-67 index, in both cases up to 12%

Fig. 2

Relative prevalence of Ki-67 pattern (homogeneous versus heterogeneous distribution) in lung carcinoids. A homogeneous Ki-67 labelling pattern was significantly more represented in the low (<4%) Ki-67 subgroup (86%, p < 0.02, Chi square test) and in the intermediate (4–9%) Ki-67 subgroup (80%, p < 0.02, Chi square test), whereas a heterogeneous Ki-67 pattern was preferentially seen in the subgroup of carcinoids with a proliferation index of ≥10% (38%, p < 0.02, Chi square test)

Fig. 3

Kaplan-Meier survival curves in lung carcinoids segregated according to histological type (a) and Ki-67 > 4% (b). a Patients with a TC diagnosis had an excellent overall survival, which was significantly better compared to patients affected by ACs (p < 0.0001). b The 4% cut-off on the whole tumour cohort effectively split patients into worse categories for tumours displaying Ki-67 higher than 4% (p = 0.003)

Fig. 4

Kaplan-Meier survival curves in lung carcinoids segregated according to distinct Ki-67 thresholds (4 and 10%). Analysis of patients’ outcome in terms of overall survival (a) and time to progression (b) in the whole cohort of lung carcinoids stratified according to 4 and 10% cut-offs. Of note, the difference in terms of outcome between the categories of <4 and 4–9% was not statistically significant for either overall survival or time to progression (p = 0.40 and p = 0.32, respectively). c Subgroup analysis restricted to atypical carcinoids segregated according to the 10% cut-off, which significantly predicted the ultimate outcome of patients (p = 0.03) even in the sole AC histological type