Autofluorescence and white light bronchoscopy for the detection of tumor margins of non-small cell lung cancer after neoadjuvant immunotherapy: a retrospective observational study.

Introduction
Globally, lung cancer remains the most common cause of cancer-related death. Non-small cell lung cancer (NSCLC) predominates in this category, representing 80–85% of cases (1). The 5-year survival rate for individuals diagnosed with NSCLC is currently at 15.9% (2). In contemporary medical practice, bronchoscopy has become a fundamental technique for the identification and therapy of NSCLC.
Autofluorescence bronchoscopy (AFB) is an established bronchoscopy technique that has been developed and refined since the late 1990s, utilizing specialized equipment to excite and identify cellular autofluorescence combined with computerized image analysis. Enhanced by camera and computer information processing, the normal trachea and bronchial mucosa tissues appear green under AFB, while atypical hyperplasia and cancer tissues appear reddish brown, pink, or red (3). Several studies have validated significant enhancement in the identification rate of precancerous lesions, carcinoma in situ, and invasive carcinoma through the utilization of AFB, which is crucial for early lung cancer diagnosis and enhancing sensitivity for precancerous lesions detection (4-12). Preoperative AFB facilitates the assessment of the dimensions and margin boundaries of early central lung carcinomas (13,14). AFB indications also extend to postoperative surveillance following curative surgical resection for NSCLC (15).
Clinical treatment for NSCLC mainly includes platinum-based chemotherapy, molecular targeted therapy, radiotherapy, cellular biotherapy, and surgical excision (16,17). Recently, the clinical addition of neoadjuvant immunotherapy has reduced the mortality rate of lung cancer. Neoadjuvant immunotherapy is a novel treatment approach in antitumor immune therapy for patients scheduled for surgery. It combines chemotherapy with immune checkpoint inhibitors, which are usually given in 2–3 cycles. However, the value of AFB in the evaluation of tumor margins of NCLSC after neoadjuvant immunotherapy has not been discussed in prior studies.
Here, we conducted a retrospective observational analysis among lung cancer patients who received neoadjuvant immunotherapy to further investigate the value of AFB in the preoperative evaluation of tumor margins. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1653/rc).

Methods

Patient selection
This was a single-center, retrospective observational case series. We analyzed the clinical data of 235 individuals diagnosed with lung cancer who received AFB at the Department of Respiratory Medicine, Second Affiliated Hospital of Zhejiang University School of Medicine, between July 1, 2022, and June 13, 2024. Ninety-two of them had received both AFB and pathological biopsy. There were 59 non-stage IV patients among these 92 patients. Finally, of these 59 patients, 22 patients who had undergone 2–3 cycles of neoadjuvant immunotherapy were included in this study. See Figure 1 for details.
Patients were all evaluated preoperatively, requiring bronchoscopic assessment to clarify the extent of lesion involvement. All patients met the basic criteria for bronchoscopy as follows: (I) no significant abnormalities in the electrocardiogram or abnormalities but stable cardiac lesions without special treatment; (II) informed consent and signed consent forms for bronchoscopy and biopsy before surgery; (III) all patients had 2–3 neoadjuvant chemotherapy sessions combined with immunotherapy; (IV) those who could tolerate anesthesia and bronchoscopy.

Bronchoscopic examination
All patients were preoperatively assessed by both white light bronchoscopy (WLB) and AFB by BF-F260 (Olympus, Japan) after 2 to 3 cycles of neoadjuvant immunotherapy. The bronchoscopic examination was conducted by physicians at our hospital who had received training in fluorescence bronchoscopy and possessed over 3 years of experience in performing bronchoscopy. The sequence of examination was WLB first, then switching to AFB mode. Repeating the above examination steps, switch to white light after finding abnormalities, and perform the biopsy under WLB.
The rules for the biopsy of lesions by endoscopists are as follows: lesions visible under bronchoscopy are defined as the primary central lesion and peripheral mucosal infiltrative lesion. The neoplastic obstruction of the trachea or bronchus is classified as the primary central lesion. Concurrently, lesions in the surrounding mucosa of the obstructed bronchus, which appear stained under AFB but normal under WLB, are classified as peripheral mucosal infiltrative lesions. To identify the tumor margins of lung squamous carcinoma, every peripheral mucosal lesion was taken for biopsy. If no peripheral mucosal lesion was found, the central primary lesion was taken for biopsy. Two to three pieces of tissue were taken from each lesion for biopsy.

Pathological evaluations
Histopathologists with over 10 years of experience in the Second Affiliated Hospital of Zhejiang University School of Medicine conducted the diagnosis of lung cancer using histological and cytological samples. The assessment depended upon hematoxylin-eosin staining and immunostaining to delineate histological type.

Statistical analysis
Continuous variables are presented as means and categorical variables as frequencies or percentages (n, %). We collected the results of AFB manifestations and pathological biopsies from 22 patients and analyzed the concordance rate between positive AFB results and pathological biopsy results. All statistical analyses were conducted utilizing the software SPSS 26.0.

Ethical Statement
The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board for Human Studies of the Second Affiliated Hospital of Zhejiang University School of Medicine (No. 2023LSYD1245) and individual consent for this retrospective analysis was waived.

Results

Patient characteristics
Clinical characteristics of 22 patients finally involved including gender, age, disease diagnosis, and disease stage are presented in Table 1. All 22 patients were males (100%), aged 64.8 years on average. Based on pathological biopsies, 21 cases were confirmed as squamous carcinoma (stage IB–IIIC), while 1 case was confirmed as adenocarcinoma (stage IIB). The expression of programmed death-ligand 1 (PD-L1) was tested in pretreatment biopsy samples from 8 patients. Four of them (50%) were positive. All 22 patients were considered by the thoracic surgery consultation to be in need of neoadjuvant treatment to re-evaluate the scope and indications for surgery. Ultimately, 13 patients underwent radical resection, whereas the remaining 9 did not proceed with surgery due to: insufficient downstaging precluding R0 resection (n=3), failure to meet predefined physiological operability criteria (n=2), or patient refusal following comprehensive counseling (n=4). Specifically, the surgical procedures comprised lobectomy (n=9), sleeve resection (n=3), and pneumonectomy (n=1). Postoperative pathological results indicated that 4 patients achieved major pathologic response (MPR) and 5 achieved pathologic complete response (pCR). The remaining 4 patients exhibited viable tumors of 10% or greater. See Table 1 for details.

AFB manifestations after neoadjuvant immunotherapy
After 2 to 3 cycles of neoadjuvant immunotherapy, the patients underwent a second AFB and pathological biopsy. All 22 patients in this post-neoadjuvant immunotherapy cohort showed abnormal pink or red lesions under AFB. After aspiration of secretions or a minor amount of congestion, these positive regions were more easily distinguished from the normal mucosa showing green color. Two typical patients’ images of bronchoscopies are presented in Figure 2. Their images of pathological biopsies are presented in Figure 3.
The details of the second AFB were: among 22 patients, 20 patients had both peripheral mucosal lesion and central primary lesion, and 2 patients only had central primary lesion. Localizations of the pink-stained sites under the AFB and the average number of biopsy blocks that were taken for biopsies are presented in Table 1.

The degree of conformity of AFB with pathological biopsy
The number of cases in which the corresponding site was biopsied accounted for 100% of the total. Pathological results of 20 peripheral mucosal infiltrative lesions were: bronchial mucosa in 8 cases (40%), extruded cells (lymphocytes) in 3 cases (15%), and chronic inflammation of the mucosa in 9 cases (45%). Pathological results of 2 central primary lesions were: chronic inflammation of the mucosa in 2 cases (100%). In summary, the concordance rate between positive AFB results and pathological biopsy results was 0%. In addition, among 13 patients who underwent complete tumor resection, MPR occurred in 4 patients, and pCR occurred in 5 patients. See Table 2 for details.

Discussion
In this retrospective clinical study, we found all positive AFB results but negative pathological biopsy results in lung cancer patients who had received neoadjuvant immunotherapy. Until now, a considerable body of research has indicated that AFB, when used in combination with WLB, serves as a valuable adjunct for identifying and characterizing central airway lesions, particularly in the detection of preinvasive squamous cell carcinoma (18). Prior studies predominantly examined the effectiveness of AFB in diagnosing early-stage lung cancer. The meta-analysis carried out by Sun et al. showed that when WLB and AFB were combined, the pooled sensitivity for identifying pre-invasive lesions was 85% and the pooled specificity was 61%, which were significantly higher than when using WLB alone (19). However, the value of AFB in patients after neoadjuvant immunotherapy has never been reported. We found positive AFB results but negative pathological biopsy results in these patients. Our study is the first to report the poor value of AFB in evaluating the tumor margins of lung cancer after neoadjuvant chemotherapy combined with immunotherapy, which contributes to the preoperative evaluation of such patients.
Unlike previous studies, our current investigation was carried out on individuals who had undergone neoadjuvant immunotherapy and were definitively diagnosed with lung carcinoma. AFB evaluation and pathological biopsy were mostly performed on the mucosal infiltrative lesions to evaluate the tumor margins of lung carcinomas. For resectable NSCLC, neoadjuvant immunotherapy can help decrease tumor size, which in turn raises R0 resection rates, and halts the progression of the disease by treating micrometastases sooner (20). Primary tumors (PTs) in NSCLC exhibit a positive pathological response to neoadjuvant immunotherapy. The absence of malignant tumor cells in surgically resected tissue is defined as pCR, whereas MPR represents a substantial reduction in viable tumor to less than 10% of its original volume. The CheckMate 816 trial proved the advantage of neoadjuvant immunotherapy over neoadjuvant chemotherapy in locally advanced NSCLC. Relevant trials have revealed unprecedentedly elevated pathological response rates with immunotherapy, exhibiting ranges of MPR rates from 18% to 83% and pCR rates spanning from 4.9% to 63% (21-26). Thus, neoadjuvant immunotherapy leading to pCR and MRP may be one of the reasons why AFB showed positive results while the pathological biopsy results were negative in the 22 cases in our study. Among 4 patients with PD-L1-negative tumors, only one patient achieved pCR. Conversely, among 4 patients with PD-L1-positive tumors, two achieved MPR, one achieved pCR, and the remaining patient declined surgery. A meta-analysis (27) indicates that a more frequent occurrence of MPR and pCR is correlated with PD-L1 expression at or above 1% compared to PD-L1 expression below 1%, which is consistent with our research.
Another reason for the false positive AFB results in this study is that AFB sometimes encounters challenges in differentiating lung cancer tissue from other epithelial changes like metaplasia, hyperplasia, and inflammation (28-30), thereby leading to unnecessary biopsies. Besides, predicting the pathological diagnosis solely based on the observed autofluorescence grade remains a complex task, as research has demonstrated instances where lesions classified as class III under AFB were subsequently identified as inflammatory lesions on biopsy (31,32). Still, studies on this phenomenon are yet to be conducted in clinical studies, and the reasons for this, as well as clinical evidence, need to be supported by larger sample sizes and more relevant studies.
In treatment-naive early central lung cancer patients, preoperative AFB combined with WLB facilitates tumor size assessment and margin delineation to inform surgical planning (13,14). A prospective study involving 23 patients diagnosed with radiographically occult lung carcinomas found that conducting supplementary staging with AFB and high-resolution computed tomography (HRCT) resulted in better identification of true occult tumors, influencing treatment plans for 70% of individuals (13). The study also indicated the feasibility of employing electrocautery during AFB, and that employing AFB can improve treatment effectiveness by precisely locating the tumor and facilitating comprehensive tumor margin resection. Another prospective study involving 104 patients confirmed that autofluorescence videobronchoscopy (AFI) substantially enhanced the evaluation of central lung carcinoma extension, resulting in modifications to the treatment strategy for 14.4% of the patients (33). However, we found that for mucosal infiltrative lesions in lung cancer patients after neoadjuvant immunotherapy, AFB results had a high false positive rate and poor agreement with pathological biopsy results, which may lead to over-resection by some physicians or even lost surgical opportunities due to incorrect information. Thus, our study reveals the unreliability of AFB and is valuable in determining the indication for surgery in lung cancer patients after neoadjuvant immunotherapy.
However, there are some limiting factors in this study. Firstly, owing to its retrospective nature, the research is potentially vulnerable to inherent biases and confounding variables. Second, the sample size is constrained by the inclusion of only 22 patients, which may limit the generalizability and statistical power of the findings. Third, we lack the performance of AFB in some of the 22 patients before neoadjuvant immunotherapy, which can help us better illustrate the lack of accuracy of AFB for preoperative assessment of tumor margins in such patients. Therefore, future validation of our findings necessitates prospective, multicentric studies with larger sample sizes.

Conclusions
The high false-positive rate of AFB findings in NSCLC patients following neoadjuvant immunotherapy is a notable constraint in its application to preoperative assessment of tumor margins in such patients, and a comprehensive preoperative determination in combination with pathological biopsy is recommended.

Supplementary
The article’s supplementary files as