Abstract:

Objective

To analyze the growth rate and CT morphological features and enhancement characteristics of benign lung nodules with concomitant growth， and to apply the AI-assisted diagnostic tool for facilitating the differential diagnosis of benign and malignant lung nodules.

Methods

From January 2016 to October 2023， 22 benign pulmonary nodules in 20 cases in the Fifth Affiliated Hospital of Sun Yat-sen University were retrospectively analyzed. The growth rate and volume doubling time （VDT） of nodules were calculated. The edge and internal imaging features and enhancement features of nodules on the initial CT and the last CT were recorded. The risk of nodules was judged by AI-assisted diagnostic tools for pulmonary nodules.

Results

Eighteen cases received surgical resection and two cases underwent percutaneous biopsy，which leading to the diagnosis of cryptococcosis in 1 case， tuberculous granuloma in 3 cases， inflammatory pseudotumor in 2 cases， nodular alveolar proteinosis in 1 case， chondromatous hamartoma in 1 case， fibrotic and necrotic nodule in 1 case and granuloma in 11 cases. The interval between the initial and last CT scans of 22 pulmonary nodules was 91～2 175 d， with a median of 419.50（298.25， 728.50） d， and the growth rate of pulmonary nodules was 16.67%～566.67%， with a median of 114.58%（26.07%， 217.13%）， and the VDT was 44.73～991.67 d， with an average of （279.55±236.06） d. At the last CT scan， there were 18 solid， 3 partially solid and 1 non-solid nodules， round and irregular lesions in 9， 13 nodules respectively， clear and blurred boundaries lesions in 18， 4 nodules respectively， spiculation， lobulation and pleural tag in 5， 7 and 10 nodules respectively， and bronchogram， bubble sign， cavity， and calcification in 2， 1， 1， and 6 nodules respectively，there are 2 nodules with satellite lesions. The mean net enhancement value of 7 nodules in the last enhanced CT scan was less than ＜20 HU. 12 nodules of the last CT scan were analyzed by AI， of which 3 were classified as high risk and 9 as low risk.

Conclusion

Some benign pulmonary nodules are similar to lung cancer in growth speed and morphology. CT enhancement and AI-assisted diagnostic tools for pulmonary nodules are helpful for the diagnosis of benign nodules， and CT-guided percutaneous biopsy may be helpful to avoid unnecessary surgery.

Key words: Pulmonary nodules, Growth, Benign, CT, Artificial intelligence