Differential Diagnosis of Pulmonary Fungal Infection and Lung Cancer via 18F-FDG PET/CT: A Retrospective Study

Background: There is limited evidence regarding the 18 F-uorodeoxyglucose positron emission tomography/computed tomography ( 18 F-FDG PET/CT) characteristics of lung fungal (LF) infections with nodules or masses, which are often misdiagnosed as lung cancer (LC) with indications for surgery. We aimed to investigate the PET/CT ndings of LF infections with nodules in comparison to those of LC and clarify the diagnostic value of 18 F-FDG PET/CT in the differential diagnosis of LF infections. Methods: We enrolled 21 patients who presented with pulmonary nodules or masses on CT, were diagnosed with LF infections, and underwent PET/CT as the LF group and randomly selected 42 patients with LC diagnosed by pathology as the LC group. Clinical and PET/CT imaging data were statistically analyzed. Results: LC was the most common misdiagnosed disease in the LF group (52.38%). There were no signicant differences in lung imaging features between the two groups. The levels of white blood cells, neutrophils, and IgG and the positive rates for fungal antigen test in the LF group were signicantly higher than those in the LC group (P<0.05). Lung masses larger than 3 cm were more common in the LC group (P<0.05). Overall, 80.95% (17/21) of patients in the LF group showed increased 18 F-FDG uptake. There were no signicant between-group differences in the maximal standardized uptake value (SUVmax, 8.20 [2.70, 12.95] vs. 8.80 [7.00, 12.38]). In the LF group, eight, ve, and eight patients had cryptococcal, Aspergillus, and Talaromyces marneffei infections, respectively, with no signicant difference in SUVmax among them (5.10 [1.70, 14.40] vs. 8.20 [1.50, 8.20] vs. 8.50 [5.10, 11.30]). Conclusions: Both LF infection and LC can present with increased 18 F-FDG uptake on PET/CT. Thus,

presenting with pulmonary masses should also be suspected to have fungal infection, even those with an increased SUVmax and simultaneous lymph node and bone involvement; particular attention is needed for patients with abnormal in ammation indexes and fungal antigen test. We should be emphasized preoperative pathological examination and fungal etiology.

Background
Positron emission tomography/computed tomography (PET/CT) can combine function, metabolism, and anatomy together. Malignant tumors show an abnormal concentration of radioactivity due to increased uptake of 18 F-uorodeoxyglucose ( 18 F-FDG); this is useful for distinguishing benign lesions from malignant lesions [1] . However, the widely used imaging agent 18 F-FDG is non-speci c, and high 18 F-FDG uptake can also occur in some fungal infections, which can then be easily misdiagnosed as malignant tumors [2] . In patients without risk factors for fungal infections, who are clinically asymptomatic, or have mild symptoms, solitary nodules or masses on lung CT and high 18 F-FDG uptake on PET/CT are often used as indicators for surgery. At present, only case reports and few small-scale prospective or retrospective studies have explored the characteristics of fungal infections presenting with pulmonary nodules or masses on 18 F-FDG PET/CT [3][4][5] . These infections are easily misdiagnosed as lung cancer (LC) with surgical recommendations. We aimed to compare PET/CT ndings between pulmonary fungal infections with nodules or masses and LC to clarify the diagnostic value of 18 F-FDG PET/CT in the clinical differential diagnosis of pulmonary fungal infections.

Study design and population
This study aimed to investigate the PET/CT ndings of LF infections with pulmonary nodules or mass shadows in comparison with those of LC and clarify the diagnostic value of 18

Imaging Protocol
Imaging was performed using the PET/CT instrument (Siemens Biograph Sensation 16, Germany) and with 18 F-FDG as imaging agent produced by the PET/CT Center of the First A liated Hospital of Guangxi Medical University. The agent had a radiochemical purity of more than 99% and was quali ed using aseptic, non-thermal, and endotoxin tests. The quality of the product met the requirements for clinical research. 18 F-FDG PET/CT was performed in line with established international guidelines for infection imaging [6][7] . All patients fasted for at least 6 h, and their blood glucose levels were in the range of 3.89-6.10 mmol/L before injection of 18 F-FDG (3.7-7.4 MBq/kg). Brain imaging was performed after 30 min, spanning the trunk from the skull base to the middle and upper femur, after emptying the bladder. The same performed range of PET examination was used for brain imaging (one bed, 8-10 min) and trunk imaging (6-7 beds, 3 min/bed). The CT parameters were as follows: pitch, 1.25; layer thickness, 5 mm; tube voltage, 120 kV; and tube current, 140 mA. The collection mode was 3D. The data were uploaded to the workstation, and coronal, sagittal, and cross-sectional fusion images acquired by CT, PET, and PET/CT were obtained.

Image Analysis
PET/CT and CT images obtained from the same machine were evaluated via visual analysis and semiquantitative maximal standardized uptake value (SUVmax) assessment. Visual analysis was performed by at least two experienced PET/CT physicians. Disagreements were resolved by consulting a radiologist.
Semi-quantitative analysis was conducted based on the concentration of 18 F-FDG in the SUV focus, and the SUVmax was automatically calculated using a computer according to the SUV focus shape.

Statistical analysis
Normally distributed measurement data were expressed as the mean ± standard deviation (SD), while non-normally distributed data were expressed as the median and interquartile range (IQR). Categorical variables were expressed as the number and percentage, and between-group differences were evaluated using the independent sample t-test, Kruskal-Wallis test, and Mann-Whitney U test. Meanwhile, betweengroup differences in categorical variables were evaluated using the chi-squared test or Fisher exact test.
All statistical analyses were performed using the Statistical Package for the Social Sciences (Windows version 25.0; SPSS Inc., Chicago, IL, USA). A two-tailed P value of < 0.05 was considered signi cant.

Patient characteristics
There were no signi cant differences in sex, age, and underlying diseases between the two groups. In the LF group, two patients had diabetes; two, long-term corticosteroid use; and one, acquired immunode ciency syndrome (AIDS). In the LC group, two patients had diabetes. Overall, 52.38% (11/21) of patients in the LF group were misdiagnosed with LC; 38.09% (8/21), tuberculosis; and 4.76% (1/21), a mediastinal tumor (Table 1).

Clinical Features And Pathological / Etiological Sources
There was no signi cant difference in systemic symptoms (including fever, emaciation, debilitation, bone pain, and neck mass) and respiratory symptoms (cough, expectoration, anhelation, chest pain, and chest distress) between the two groups (P = 0.484). For the classi cation of fungal pathogens, eight, eight, and ve patients in the LF group had Talaromyces marneffei, cryptococcal, and Aspergillus infections, respectively. For pathological types, 29, 6, and 7 patients in the LC group had adenocarcinoma, squamous cell carcinoma, and small cell LC, respectively. Overall, 54.54% (6/11) of patients in the LF group were misdiagnosed to have LC and underwent surgery. Of them, ve patients underwent thoracoscopic wedge pneumonectomy and one underwent right upper lobectomy plus right middle lobectomy. Diagnosis was con rmed via surgery, biopsy, and culture in 6/21 (28.57%), 7/21 (33.33%), and 8/21 (38.10%) patients in the LF group, respectively. Meanwhile, in the LC group, diagnosis was con rmed via surgery in 24/42 (57.14%) patients; biopsy in 14/42 (33.33%) patients; and ne-needle aspiration cytology in 4/42 (9.52%) patients.

Laboratory Findings
The levels of white blood cells (WBC), neutrophils, and IgG and positive rates for fungal antigen tests (G test, GM test, and cryptococcal latex agglutination test) were signi cantly higher in the LF group than in the LC group (P < 0.05). However, there was no signi cant difference in the level of lymphocytes, CD4 + T cells, CD8 + T cells, IgA, and IgM between the two groups ( Table 2).

High-resolution Computed Tomography Features Of The Chest
There was no signi cant difference in the distribution of localized pulmonary lesions, disseminated lesions, and pulmonary imaging features (e.g., spiculation sign, necrotic cavity, pleural traction sign, and pleural effusion) between the LF and LC groups. Both lymph node and bone involvement were observed in the two groups. Lung masses larger than 3 cm were more common in the LC group (P < 0. 05) ( Table 3).

Discussion
The diagnostic value of PET/CT in malignant tumors is recognized. Tumors with increased 18 F-FDG uptake and high SUVmax are considered malignant and recommended for early surgical resection.
However, high 18 F-FDG uptake is non-speci c and can occur in other conditions. In this study, 52.38% of patients with lung fungi were misdiagnosed with malignant tumors, and 54.54% of these patients underwent surgical treatment and were eventually pathologically diagnosed with fungal infection. This indicates that LF infection is easily misdiagnosed as LC. This not only increases the economic burden of patients, but also the related complications and sequelae of lobectomy, which have a negative impact on the patients' quality of life.
Case reports have shown that patients who clinically presented with LC were eventually con rmed to have fungal infections such as Cryptococcal infection, invasive Aspergillus infection, or tuberculosis and granulomatous lesions [8][9][10][11][12][13][14] . Aside from malignant tumors, other benign lesions or certain infections can show increased glucose metabolism and present with similar manifestations. Patients with pulmonary fungal infection without underlying disease and normal immune function often show isolated nodules or masses with mild or no clinical symptoms. As such, it is di cult to distinguish these masses from LC lesions using routine CT examinations. These lesions that present with increased SUVmax on PET/CT are often misdiagnosed as LC and recommend for surgical treatment. This leads to the spread of the fungal infection, such as cryptococcal meningitis caused by the deterioration of lung cryptococci [15] . Therefore, it is important to compare and analyze common clinical fungal infections with the PET/CT manifestations of LC to improve the accuracy of diagnosis.
The risk factors for LF infections include immunode ciency, agranulocytosis, diabetes, long-term corticosteroid use, and immunosuppression. The host factor is a key factor in the hierarchical diagnosis of invasive LF infection [16][17] . However, increasing evidence indicates that in hosts without any underlying disease and with normal immune function, LF infection often presents as a solitary nodule or masse, and the clinical symptoms are mild. In this study, 5 of the 21 patients in the LF group were immunocompromised (one had AIDS; two, diabetes; and two, long-term corticosteroid use). With respect to clinical symptoms and HRCT ndings, there was no signi cant difference in the location of the lesion between the LF and LC groups. However, LF infection was accompanied by high WBC counts and positivity for fungal antigen tests. Nonetheless, it was di cult to distinguish between LF infection and LC based on host factors and clinical and HRCT manifestations. This can cause a high rate of misdiagnosis and incorrect treatment.
We found no signi cant difference in the SUVmax between the LF and LC groups, clarifying the value of PET/CT in the differential diagnosis of malignant tumors and LF infections with pulmonary lesions. After entering the lungs, fungal spores that cannot be eliminated by the body are swallowed by macrophages. The fungi form lesions with histiocytic proliferation, accompanied by necrosis, in ammatory cell in ltration or granuloma in the lungs. The macrophages that engulf pathogens invade the body's mononuclear-macrophage system along with the lymphatic system and blood system organs (e.g., lymph nodes, bone marrow, blood, liver, and spleen). During granuloma formation or in ammatory activity, the levels of glycolytic enzymes in in ammatory cells (e.g., neutrophils and macrophages) and broblasts are increased, and glucose transporters (glucose transport protein, Glut) are overexpressed. Further, increased FDG penetrates the cell membrane, manifested by an increased 18 F-FDG uptake in the primary lung lesions and other involved organs such as the lymph nodes and bone marrow [18][19] . SUV is currently the most used semi-quantitative index to evaluate the degree of glucose uptake of lesions, and most researchers consider an early phase SUVmax of ≥ 2.5 to be one of the criteria for distinguishing between benign and malignant lesions [20] . However, 13 patients in our LF group had a SUVmax > 2.5, indicating that assessments based on glucose metabolism alone cannot distinguish LF infections from LC. In the LF group, 16 12 patients showed a localized increase in SUVmax (lymph nodes, n = 12; bones, n = 4), similar to LC. Therefore, whether only the primary lung lesion, or the involvement of lymph nodes and bone concurrently, cannot be used as a basis for distinguishing between benign and malignant lesions. In addition, we also found it is di cult to distinguish the different fungal infections in PET/CT examination of pulmonary lesions.
This study has some limitations. First, the number of cases was limited; thus large, multi-center research is needed for further investigating the PET/CT ndings of pulmonary fungal infections with nodules or masses. Secondly, this was a retrospective study, more studies are needed to address key questions regarding the use of prospective surveillance and optimal treatment strategies.

Conclusions
Lung lesions presenting with characteristics highly similar to those of malignant tumors on PET/CT should be carefully evaluated to rule out other conditions. A combination of in ammatory indices such as WBC counts, erythrocyte sedimentation rate, and C-reactive protein, should be included in the assessment. Concurrently, a possible etiological basis for fungal culture, B-ultrasound, or CT-guided puncture biopsy and beroptic bronchoscopy before surgery should be established to reduce misdiagnosis and unnecessary surgeries. A lung mass larger than 3 cm, increased 18  Availability of data and materials: All data generated or analyzed in this study are included in this published article