This retrospective study was conducted from January 2022 to December 2022 at Wuhan Pulmonary Hospital. The study included patients diagnosed with inactive PTB (post-tuberculosis lung disease) complicated by CPA. All PTB cases were confirmed to have achieved clinical cure through standard anti-tuberculosis treatment, with a diagnosis of inactive PTB made at least 2 years prior to the CPA diagnosis; no evidence of active PTB lesions was observed on chest imaging, and sputum smear/culture results were negative at the time of CPA diagnosis. Patients with active PTB or concurrent active PTB and CPA were excluded. Among the included patients, 100% had a history of PTB (post-tuberculosis lung disease), and 0% had active PTB at the time of CPA diagnosis. All CPA diagnoses were independently reviewed and confirmed by at least two senior respiratory medicine specialists.

PTB was diagnosed using sputum smear microscopy, culture, or GeneXpert detection of Mycobacterium tuberculosis from clinical specimens. CPA diagnosis and classification were based on criteria outlined in the European guidelines (Denning et al., 2016), and included the following: (1) Chest computed tomography (CT) findings consistent with CPA imaging features. (2) Microbiological evidence or immunological evidence of Aspergillus infection, defined by at least one of the following: (a) positive fungal culture for Aspergillus on sputum or bronchoscopic lavage fluid (BALF); (b) positive nucleic acid for Aspergillus on sputum or BALF; (c) positive galactomannan (GM) antigen test on BALF; (d) positive serum Aspergillus IgG; (e) Aspergillus infection on histopathology of the lung. (3) Clinical, radiological, and microbiological features persisting for at least three months.

Exclusion criteria included: (1) Absence of a prior history of PTB (excluding patients with no PTB history). (2) Active PTB or concurrent active PTB and CPA. (3) Diagnosis of invasive pulmonary aspergillosis. (4) Diagnosis of allergic bronchopulmonary aspergillosis. (5) Presence of non-Aspergillus fungal infections. (6) Clinically suspected but unconfirmed Aspergillus infection.

1. Stable: 1) Following surgery and/or antifungal therapy for ≥3 months, respiratory symptoms were significantly relieved, respiratory fungal cultures were negative, and chest imaging showed improvement. Specifically, non-cavitary lesions were reduced by ≥50% (persistently stable dense nodules ≤1 cm were excluded), cavity wall or adjacent pleural thickness (baseline ≥2 mm) was reduced by ≥20%, and fungal balls or bands had resolved. No imaging deterioration was observed in other regions, and no new lesions suspicious for CPA appeared within 6 months postoperatively. 2) Following surgery and/or antifungal therapy for ≥3 months, respiratory symptoms improved, and chest imaging remained stable without progression. 3) No antifungal treatment was performed, and the symptoms and imaging were stable within 6 months, and the microbiological evidence of respiratory tract Aspergillosis was negative.

2. Worsened: 1) In patients receiving surgery and/or antifungal therapy, clinical symptoms worsened, chest imaging showed deterioration or respiratory specimens remained positive for Aspergillus at 3 and 6 months, with other causes of deterioration excluded. 2) clinical symptoms or chest imaging findings worsened by 6 months in patients not treated with antifungals.

3. Death: Mortality was defined as all-cause death occurring within one year of follow-up.

Sputum samples were inoculated onto Sabouraud dextrose agar and BALF samples onto Sabouraud glucose agar, followed by incubation at 25°C and 35°C for 2–7 days. Fungal colonies were examined daily. The presence of characteristic Aspergillus structures, such as conidiophores and foot cells, indicated Aspergillus growth. Species identification was primarily based on macroscopic colony morphology and microscopic characteristics. Nucleic acid detection of Aspergillus was performed using real-time PCR (Dilan, China), with a cycle threshold (Ct) value ≤36 considered positive. GM antigen testing in serum or BALF was performed using enzyme-linked immunosorbent assay (ELISA; Bio-Rad Laboratories, CA, USA) according to the manufacturer’s instructions. For diagnostic interpretation, a GM index (GMI) ≥0.5 was defined as positive for serum, while a more specific threshold of GMI ≥0.8 was adopted for BALF, in accordance with current guidelines (Salzer et al., 2018; de Oliveira et al., 2023). Aspergillus-specific IgG antibody testing was conducted using ELISA (Dynamiker, China). It is important to note that this assay uses Aspergillus galactomannan as the plate-coating antigen. Based on the manufacturer’s instructions and for the purpose of this study, a value ≥80 AU/mL was considered positive (with the equivocal range of 80–120 AU/mL included in the positive group for analysis).

Quantitative variables were expressed as mean ± standard deviation (SD) for normally distributed data or as median with interquartile range (IQR) for non-normally distributed data. Categorical variables were presented as counts and percentages. Comparisons between groups were performed using the Student’s t-test for normally distributed variables and the Mann–Whitney U test for non-normally distributed variables. Significance was set at P <0.05. GraphPadPrism 9 (GraphPad Software, La Jolla, CA, USA) was utilized to analyze data.

A total of 220 PTB patients with CPA were included in the study. The cohort comprised 31 with SA, 120 with CCPA, 39 with CFPA, 25 with AN, and 5 with SAIA. Patient general characteristics are summarized in Table 1. The study cohort had a mean age of 56.7 ± 14.7 years (range: 21–84 years), with a significant male predominance (141 males, 79 females). Male patients were the majority in the CFPA, CCPA, and SA subgroups. Patients with CFPA and CCPA were older compared to those in other subtypes.

Most patients presented with pre-existing pulmonary or systemic comorbidities in addition to PTB. The most prevalent conditions were bronchial tuberculosis (n = 56, 25.5%), chronic obstructive pulmonary disease (COPD, n = 49, 22.3%), and diabetes mellitus (n = 44, 20.0%), followed by bronchiectasis (n = 31, 14.4%), interstitial lung disease (n = 4, 1.8%), and lung tumor (0.9%, n = 2). A diagnosis of bronchial tuberculosis was based on either a documented prior history (self-reported or bronchoscopy-confirmed records) or characteristic chest CT findings such as bronchial wall thickening and luminal stenosis. Among subgroup analyses, bronchial tuberculosis was the most common comorbidity in the AN (36.0%, n = 9) and SAIA (n = 2, 40.0%) subgroups, while COPD was frequently observed in CCPA (n = 32, 26.7%) and CFPA (n = 9, 23.1%). Diabetes mellitus was more prevalent in SA (n = 8, 25.8%) and CCPA (n = 26, 21.7%).

Ten patients had a history of prolonged glucocorticoid use, including seven with CCPA, two with SA, and one with AN. Two cases of lung cancer were identified, affecting one patient each in the CCPA and CFPA subgroups. Interstitial lung disease and pneumoconiosis were primarily observed in the SA and CCPA subgroups, while no cases of pulmonary sarcoidosis were detected.

Clinical and radiological data are summarized in Table 2. The most frequently reported symptoms were cough (n = 207, 94.1%) and sputum production (n = 183, 83.2%), followed by hemoptysis (n = 126, 57.3%) and fatigue (n = 45, 20.5%). These symptoms were most prevalent in the CCPA, CFPA, and SAIA subgroups. Hemoptysis occurred more frequently in CFPA (79.5%) and SA (71.5%) than in CCPA (55.8%). Fever was predominantly observed in AN (55.8%) and SAIA (40.0%). Notably, fever was observed in 2 of the 5 (40.0%) SAIA patients; however, given the very small sample size (n = 5) of this subgroup, this finding must be interpreted with caution and cannot reliably support comparative claims regarding fever rates across subtypes. CFPA patients exhibited significantly higher incidences of dyspnea (51.3%), chest pain (28.2%), and weight loss (53.8%) compared to other groups.

Chest CT imaging was abnormal in all enrolled patients. The lesions primarily involved the upper lung zones (85.9%), with distribution across the right lung (162 patients), left lung (132 patients), and lung field (218 patients), consistent with bilateral disease in a significant portion of the cohort. Dominant CT abnormalities included cavity formation (87.7%), intracavitary septum (75.5%), progressive pleural thickening (75.5%), fibrosis (58.6%), and emphysema (55.0%). Cavity formation was universal in SA, CCPA, and CFPA but rare in AN (n = 1) and SAIA (n = 2). CFPA showed significantly higher rates of emphysema (79.5%) and fibrosis (89.7%) than other subtypes, reflecting advanced fibrotic remodeling.

High-resolution computed tomography (HRCT) findings at the time of diagnosis are presented in Figure 1. All CT/HRCT images were independently reviewed by two board-certified radiologists with ≥5 years of experience in thoracic imaging; discrepancies were resolved by consensus with a senior thoracic radiologist (≥10 years of experience) blinded to clinical and microbiological data. Radiologic assessments and subtype definitions strictly followed the ESCMID/ERS guidelines (Denning et al., 2016), with core imaging features defining each subtype: AN presents as one or more nodules (with or without cavitation), frequently showing necrosis and mimicking radiological features of tuberculoma or lung carcinoma; CCPA, the most prevalent subtype, is characterized by one or more thin- or thick-walled pulmonary cavities, potentially containing aspergillomas or irregular intraluminal material, along with overt radiological progression (new cavities, increasing pericavitary infiltrates, or progressive fibrosis) over at least 3 months; CFPA manifests as severe fibrotic destruction involving at least two lung lobes (with focal fibrotic cavitation in a single lobe classified as CCPA), typically presenting as consolidation or large cavities surrounded by fibrosis; SA is defined by a solitary pulmonary cavity containing a fungal ball and no radiological progression over ≥3 months; and SAIA exhibits variable imaging features including cavitation, nodules, and progressive consolidation with abscess formation. CT imaging plays a critical role in the rapid evaluation of CPA, particularly in cases where microbiological confirmation is delayed. Radiological phenotypes varied across the CPA spectrum but consistently demonstrated upper lobe predominance. Overlapping radiological features were frequently observed (e.g., CCPA-AN/SAIA transitions), supporting the concept of CPA as a clinical-radiological continuum rather than a set of distinct entities consistent with its progressive pathological nature.

Laboratory and microbiological data are presented in Table 3, with subtype-specific variations illustrated in Figure 2. All 220 CPA patients had direct evidence of Aspergillus infection. Most patients exhibited normal white blood cell (WBC) counts (median 6.2×10⁹/L, IQR 4.9-7.8) and neutrophil counts (median 4.2×10⁹/L, IQR 3.1-5.6), though CFPA (median 4.7×10⁹/L) and CCPA (4.3×10⁹/L) showed significantly higher neutrophilia than AN (3.4×10⁹/L, P < 0.01 for both). Lymphocyte counts were below normal in SAIA (0.8×10⁹/L) and CFPA (0.9×10⁹/L), while AN had higher counts (1.7×10⁹/L) compared to CCPA (1.2×10⁹/L) and SA (1.2×10⁹/L), remaining within physiological limits. The low lymphocyte count in the SAIA subgroup (n = 5) is noted but requires validation in larger cohorts. Hypochromic anemia was common across the cohort, with a median hemoglobin (Hb) level of 118 g/L. Anemia was most severe in CFPA (102 g/L), while SA (123 g/L) and CCPA (117 g/L) had significantly lower Hb compared to AN (131 g/L; P < 0.05). Platelet (PLT) counts were generally within normal limits (median: 225 × 10⁹/L), though mild thrombocytosis was noted in CCPA (239 × 10⁹/L) and CFPA (247 × 10⁹/L). Hypoalbuminemia was more pronounced in CFPA (27 g/L) and CCPA (31 g/L) compared to other subtypes (P < 0.05). Inflammatory markers were elevated across the cohort. The erythrocyte sedimentation rate (ESR) showed a median value of 35 mm/h, with the highest levels observed in SAIA (66 mm/h), followed by CFPA (60 mm/h). High-sensitivity C-reactive protein (hsCRP) levels were also elevated, peaking in CFPA (30.6 mg/L) and followed by CCPA (16.8 mg/L).

Pathological confirmation was obtained using next-generation sequencing (NGS) or polymerase chain reaction (PCR) in 119 patients. Additional diagnostic methods included aspiration biopsy (n = 1), bronchoscopy (n = 6), and surgical excision (n = 43). The overall fungal smear positivity rate was 10.0% (n = 22). Culture positivity varied by subtype: CFPA (35.9%), CCPA (23.3%), SA (22.6%), SAIA (20.0%), and AN (16.0%). Serum GM index levels were predominantly within normal ranges, whereas BALF GM index more frequently exceeded diagnostic thresholds. BALF GM positivity was highest in AN (76.0%), significantly exceeding SA (41.9%), CCPA (52.5%), and SAIA (20.0%, P < 0.05). Serum Aspergillus-specific IgG was positive in 68.2% of patients, with CFPA (74.4%), CCPA (72.5%), and AN (72.0%) showing higher rates than SA (48.4%) and SAIA (20.0%). The low rates of BALF GM positivity and serum IgG positivity in the SAIA subgroup (n = 5) are presented; however, the very small sample size limits any conclusive interpretation of these findings.

Given the critical role of GM in CPA, correlations between GM levels and laboratory parameters were further analyzed (Figure 3; Supplementary Figures S1–12). Overall, serum GM in CPA patients showed a negative correlation with Hb and a positive correlation with both ESR and BALF GM. No significant associations were found between BALF GM and peripheral blood parameters, possibly due to limited systemic GM release. Subtype-specific correlation patterns revealed further distinctions. In CCPA, serum GM negatively correlated with Hb and positively correlated with BALF GM. In SA, BALF GM negatively correlated with Hb. In AN, BALF GM was positively associated with hsCRP, with no other significant correlations. In CFPA, BALF GM positively correlated with lymphocytes and PLT. Finally, only a positive correlation between BALF GM and Alb was detected in SAIA. Notably, the SAIA subgroup included only 5 patients, so this correlation should be interpreted with caution due to the small sample size.

As outlined in Table 4, antifungal therapy was administered to 90.0% (198 of 220) CPA patients. Among them, 155 received pharmacologic treatment alone, while 43 underwent combined antifungal therapy and surgical intervention. Voriconazole was the most frequently prescribed agent (n = 152, 69.1%), followed by itraconazole (n = 13, 5.9%) and amphotericin B (n = 3, 1.4%). Adverse drug reactions (ADRs) were reported in 19 patients, with the highest incidence observed in the AN subgroup (n = 4, 16%). Most patients (n = 155, 70.5%) received antifungal therapy for ≤6 months, whereas 57 patients did not undergo any antifungal treatment—primarily due to economic constraints preventing the affordability of antifungal agents or follow-up observation recommended for asymptomatic patients with radiologically stable CT lesions following clinical evaluation. Only eight patients continued treatment beyond six months, including 1 SA, 5 CCPA, and 2 CFPA cases.

Clinical response evaluation revealed that 167 patients (75.9%) achieved stable disease, while 49 (22.3%) experienced disease progression. Four patients (10.3%) in the CFPA subgroup died during follow-up, with no deaths directly attributable to CPA, massive hemoptysis, or fungal sepsis. Adjudicated causes included surgery-related severe pneumonia, advanced malignancy with severe pneumonia, drug-induced multiple organ failure, and COPD acute exacerbation with respiratory failure—consistent with the subgroup’s baseline characteristics (older age, lower hemoglobin and albumin, higher rates of COPD and malignancy). CFPA patients had the poorest overall outcomes, with a significantly lower stable disease rate (51.3%) and the highest progression rate (38.5%) among all subtypes. In contrast, AN patients showed the most favorable outcomes, with the highest stable disease rate (92.0%) and the lowest progression rate (8.0%).