All samples were obtained from the Department of Pathology, Sun Yat-sen University Cancer Center (April 2004 to June 2021). All enrolled patients had a confirmed PC diagnosis. The diagnostic criteria for the included PC patients were: sections of lung tissue were stained with GMS, PAS, and AB, and at least one staining was positive for Cryptococcus. Hematoxylin and eosin (H&E), GMS, PAS, and AB staining sections were collected from the archives of all enrolled patients, and the clinical data were extracted from medical records. The Institutional Review Board of Sun Yat-sen University Cancer Center approved this study (B2021-452-01).

The accessed PC tissue staining sections (n = 152) were prepared by trained technicians from the histology laboratory using standardized staining procedures of H&E, GMS, PAS, and AB. All sections were 4-μm thick and sequential. H&E staining section was the last wax slice. Lung tissue infected with Cryptococcus was used as a positive control, and normal human lung tissue was used as a negative control.

Routine H&E staining: the sections were stained using an H&E staining kit (Baso Diagnostics Inc., Zhuhai, China). Paraffin sections were dewaxed with xylene and gradient alcohol, and then stained with H&E. Finally, the sections were dehydrated in a gradient and sealed with a neutral resin.

Grocott methenamine silver staining: sections were stained using a GMS staining kit (Baso Diagnostics Inc., Zhuhai, China). Paraffin sections were dewaxed with xylene and gradient alcohol and then oxidized in 8% chromic acid solution for 20 min. After pretreatment with 0.5% sodium metabisulfite solution, the sections were immersed in a methenamine-silver nitrate solution preheated to 60°C and placed in an electro-thermostatic blast oven at 60°C for 40 min. The sections were subsequently washed with distilled water, treated with a 5% sodium thiosulfate solution, and counterstained with a light green solution. Finally, the sections were dehydrated in a gradient and sealed with a neutral resin.

Periodic acid-Schiff staining: sections were stained using a PAS staining kit (Baso Diagnostics Inc., Zhuhai, China). Paraffin sections were deparaffinized using xylene and gradient alcohol. After oxidation with a 1% periodic acid solution for 10 min, they were treated with Schiff reagent in the dark for 20 min. The sections were then washed under running water for 5 min and counterstained with Mayer’s hematoxylin. Finally, the sections were dehydrated in a gradient and sealed with a neutral resin.

Alcian blue staining: sections were stained using an AB staining kit (Baso Diagnostics Inc., Zhuhai, China). Paraffin sections were deparaffinized using xylene and gradient alcohol. After staining with AB (pH 2.5) solution for 20 min, the sections were stained with nuclear fast red solution for 5 min. Finally, the sections were dehydrated in a gradient and sealed with a neutral resin.

Cryptococcus-positive diagnosis with GMS, PAS, and AB staining was indicated by brown-black staining with a light green background, magenta staining with a blue nucleus, and blue staining with a red nucleus, respectively. Pathological diagnoses of all stained sections and cases were independently evaluated by two pathologists; any discrepancies were resolved by a third senior doctor. Statistical analysis was performed using SPSS software (version 25.0; SPSS, Chicago, IL, United States). The quantitative index of normal distribution was described by x̄ ± s. The classification index was expressed by case count and percentage. McNemar’s test was used to compare the sensitivity of GMS, PAS, and AB stains. P-values of <0.05 were considered statistically significant.

A total of 152 (53 women; average age, 52 years; age range, 23–81 years) patients with PC were included. Among them, 102 (67.1%) patients had no clinical symptoms and were admitted because of the presence of pulmonary nodules on physical examination. Leukocyte count, neutrophil percentage, and C-reactive protein levels were within the normal ranges in 141 (92.8%), 135 (88.8%), and 140 (92.1%) patients, respectively. This evidence suggests that in cases diagnosed with PC, infection indicators may not show a correlation, which may be related to the body’s immunity. The imaging findings of patients with PC were not specific, and multiple nodules (96 cases, 63.2%) were common, with spiculation and lobulation of suspected lung cancer in 67 (44.1%) and 78 (51.3%) cases, respectively. These findings suggest a lack of imaging or clinical presentation specificity associated with PC (Table 1).

Statistical analysis showed that the detection rates of GMS, PAS, and AB staining were 100% (152/152), 94.7% (144/152), and 81.6% (124/152), respectively, in paraffin-embedded tissue sections from 152 patients with PC. The sensitivity of GMS staining was 100%, whereas that of PAS staining was 94.7%, and the consistency between them was 94.7%; McNemar’s test showed that the difference between the two stains was statistically significant (P = 0.008). The sensitivity of AB staining was 81.6%, and the consistency between GMS and AB was 81.6%; there was a significant difference between the two methods (P < 0.001). The sensitivity of PAS was higher than that of AB, and the consistency between the two methods was 84.2%; the difference between the two methods was statistically significant (P < 0.001). These findings demonstrated that the sensitivity of GMS staining was significantly higher than those of PAS staining (P = 0.008) and AB staining (P < 0.001) (Table 2).

The positivity rates of GMS, PAS, and AB staining were mainly observed in four modes (Table 3). The most common mode was positive for GMS, PAS, and AB staining. There were 122 cases in Mode A, accounting for 80.3% of the study cohort. In these cases, H&E staining revealed morphological features of granulomatous inflammation in the lung tissue. Hyperplastic epithelioid cells and multinucleate giant cells were observed, with diaphanous round or oval structures in the cytoplasm, suspected to be Cryptococcus spores (Figure 1A), and infiltration of interstitial lymphocytes and plasma cells. Occasionally, necrosis was observed in the lung tissue. GMS staining revealed a brown-black color that contrasted with a bright green background. Cryptococcus presented clearly as round or oval structures with different sizes. It had a thick mucinous capsule and narrow-based budding (Figure 1B). The background was clean, and Cryptococcus could be quickly recognized at low- or medium-power magnification. Meanwhile, PAS and AB stained Cryptococcus with magenta and blue of varied intensity, respectively; some spores stained normal magenta and blue, and some were lightly stained (Figures 1C,D). Moreover, under the same field of view, PAS and AB staining showed lower Cryptococcus density than did GMS staining.

In Mode B, 22 (14.5%) cases were positive for GMS and PAS staining, and negative for AB staining. In these 22 cases, H & E staining revealed circular structures resembling Cryptococcus spores (Figure 2A), and, Cryptococcus was clearly highlighted in brown-black by GMS staining with a clean background (Figure 2B). Although Cryptococcus can be found in some multinucleated giant cells, PAS staining requires careful identification of magenta circular structures with obvious positive characteristics under high-power magnification, which is time-consuming and laborious (Figure 2C). It was difficult to identify the morphology of blue Cryptococcus with AB staining at low- or high-power magnification. Although there were similar structures to Cryptococcus, the final diagnosis involved non-specific staining after careful identification (Figure 2D). In contrast, there were 2 (1.3%) cases in which GMS and AB staining were positive and PAS was negative in Mode C. H&E staining revealed circular structures resembling Cryptococcus spores (Figure 3A), and GMS staining highlighted brown-black Cryptococcus spores (Figure 3B). The circular structures suspected to be Cryptococcus in the PAS staining sections of the two cases were colorless or reddish, with no obvious magenta circular structures (Figure 3C). AB staining revealed Cryptococcus blue within the multinucleated giant cells, with interfering blue acid mucosubstances (Figure 3D). In Mode D, 6 (3.9%) cases were positively diagnosed with GMS staining only. In this type of cases, H&E staining and GMS staining tissue morphology were similar to corresponding Mode A, Mode B and Mode C (Figures 4A,B). The magenta circular structure of the suspected Cryptococcus with PAS staining was covered with non-specific staining, making it difficult to identify magenta Cryptococcus (Figure 4C). AB staining did not reveal Cryptococcus blue (Figure 4D). These results indicated that GMS staining was superior to PAS and AB staining in terms of quality, stability, and identifiability.

In PC sections stained with GMS, Cryptococcus was brown-black and prominent against the bright green clean background. The size, structure, and narrow-based budding were observed at low- or medium-power magnification (Figures 5A,C). However, there was interference of red coloration revealing neutral mucosubstances, glycogen, and some connective fiber tissues in PAS staining. Specifically, when Cryptococcus accumulated in the necrotic foci, magenta Cryptococcus was similar in color to the necrotic foci and was not easily distinguished (Figures 5B,D). However, when the staining background had a large number of normal human cells, it interfered with Cryptococcus observation, which usually requires high magnification. Similarly, AB staining interfered with the blue acid mucosubstances. Non-specific PAS and AB staining resulted in a structural pattern resembling that of Cryptococcus. In PAS staining, this structural pattern was red, round-like, with a colorless middle and a reddish periphery (Figure 2C), whereas for AB staining, the same structural pattern was formed with a colorless middle and blue outer ring (Figure 2D). This structural pattern, comparable to that of Cryptococcus, can be missed even at high-power magnification. Comparative analysis showed that the identifiability of GMS was significantly better than those of PAS and AB staining, and that GMS was optimal for diagnosing PC.