A Retrospective Study on Clinical Signi cance of Imaging And Endobronchial Features In Pulmonary Lymphoma Diagnosed By Bronchoscopy: When Should We Perform Bronchoscopy?


 Background: Pulmonary lymphoma is rare while the radiographic lesions are nonspecific. The use of bronchoscopy, is controversial, and may be of limited diagnostic value for pulmonary lymphoma diagnosis.Methods: Cases of pulmonary lymphoma diagnosed by bronchoscopy were retrospectively reviewed from January 2010 till December 2020. Clinico-radiological records of these cases were retrieved and bronchoscopy was performed with histopathological evaluation of the biopsy specimens. Results: Of the 41 cases included in our study, 20 were primary pulmonary lymphoma. The majority of subtypes were diffuse large B-cell lymphoma(43.9%) and extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue lymphoma(29.3%), with high percentage of underlying chronic disease. The common Computed Tomography(CT) abnormalities were mass(43.9%), consolidation(24.4%) or ground-glass opacities(24.4%). To evaluate the pulmonary lesions distribution, 41.5% patients had segmental lesions, 29.3% presences of diffuse lesions, 17.1% of local lesions and 12.2% of hilar lesions, with the maximal value of standard uptake value (SUVmax) median value of 9.6. At bronchoscopy, 41.5% lesions were identified as focal solitary mass, 22% were normal, 17.1% were mucosae asperity or edema, 9.8% were diffuse airway submucosal nodules or thickening and 9.8% were bronchiole stenosis, while certain endobronchial patterns may correspond to specific imaging features. Conclusion: The imaging features and endobronchial patterns of pulmonary lymphoma are characteristic. Bronchoscopy is still a useful tool, that is minimally invasive, to diagnose pulmonary lymphoma, while accurate analysis of the clinico-radiological records is essential before such procedure.


Background
Pulmonary lymphoma is involved in < 10% of Hodgkin lymphomas and < 5% of non-Hodgkin lymphomas.
The radiographic patterns in the lung include nodular, patchy pneumonic in ltrates, mass, lymphomatous lesions, or military hematogenous forms [1]. Lymphoma manifesting such as lung opacities is sometimes di cult to diagnose bronchoscopically, that Computed Tomography(CT) guided transthoracic needle biopsy or even thoracoscopy has been described as a useful modality for obtaining larger biopsy samples in diffuse parenchymal lung lesions. However, pulmonary physicians still tend to perform bronchoscopy as the rst step on patients of suspected pulmonary lymphoma with certain CT imaging features. Are there any speci c imaging features in patients diagnosed under bronchoscopy? What are their corresponding endobronchial lesions under bronchoscopy? In this study, we retrospectively analyze the clinical manifestation and features of patients bronchoscopically diagnosed as pulmonary lymphoma.

Patients
A total of 41 patients with pulmonary lymphoma diagnosed by bronchoscopy were retrospectively identi ed through key words search (bronchoscopy, lymphoma) of pathological databases at Peking Union Medical College Hospital for clinical assessment (PUMCH, Beijing, China), who were diagnosed between January 1, 2010, and December 31, 2020. The study protocol was reviewed and approved by the ethical committee of Peking Union Medical College Hospital, and written consent was obtained from all subjects.

Imaging Protocol
All patients were scanned using axial CT was performed (140 kV, 35 mA, pitch of 1:1, 5-mm thickness with 3-mm gap, 512×512 matrix, 70-cm eld of view,6msv) from the neck to the mid-thigh performed in inspiratory apnoea before the bronchoscopy. Precontrast and enhanced images were obtained in 16 patients. As previous study mentioned [2], features of the pulmonary lesions were described as localization (bilateral or unilateral), distribution (local, segmental, diffuse or hilar) and morphological pattern (consolidation, mass, nodule, or ground-glass opacities(GGO), et al). Area of consolidation was de ned as homogeneous increase in pulmonary parenchymal attenuation obscuring the vessels margins and airways walls; mass was de ned as rounded well-de ned or moderately well-de ned opacity greater than 3 cm in maximum diameter; nodule was de ned as rounded well-de ned or moderately well-de ned opacity equal to or less than 3 cm in maximum diameter; GGO was de ned as a hazy increased attenuation of lung with preservation of bronchial and vascular margins, and mosaic pattern was a condition in which there was a patchwork of regions of different lung attenuation on thin-section CT.
A total of 21 patients had undergone uorine18 2deoxy2 uoroDglucose (18FFDG) Positron Emission Computed Tomography(PET) scanning. PET/CT scans were performed on the PET/CT system (Biograph 64 TruePoint; Siemens, Knoxville, Tennessee, USA). The maximal value of standard uptake value (SUV max ) in pulmonary lesions was recorded.

Bronchoscopy
All patients underwent a diagnostic bronchoscope (Table 3). Among 41 patients, right lower lobe(24.4%), left upper lobe(19.5%) and left lower lobe(19.5%) were the most frequently examined location, as physicians prefer to perform TBLB in the lower lobe, especially patients with normal endobronchial pattern. A total of 35 pulmonary lesions were identi ed: 17 areas of focal solitary mass, 7 mucosae asperity or edema, 4 diffuse airway submucosal nodules or thickening, 4 bronchiole stenosis and 3 only lymph node enlargement( Fig. 1), while stations 4R and 7 were the most frequently examined lymph node.
One bleeding within 48 hours after neoplasm biopsy as procedure-related complication occurred in this study.

Diagnosis
Diagnosis of lymphoma was according the updated WHO classi cation of Lymphoid Malignancies by PUMCH pathology department. Some diagnosis were established by MDT (Multiple Disciplinary Team) with experienced hematopathologist with the help of immunohistochemistry or even ow cytometry [3]. While primary pulmonary lymphoma (PPL) is de ned as a lymphoma con ned to the lung with or without hilar lymph node involvement at the time of diagnosis or up to 3 months thereafter [1,4].

Statistical analysis
Statistical analyses were processed by statistical software program (PASW statistics 16; SPSS Inc., Chicago, IL., USA). Means and percentages were presented as appropriate.

Patients
All 41 patients were histologically con rmed with pulmonary lymphoma. The female: male ratio is 1:1.1, while median age is 52 years(range 2277 years). All patients underwent a diagnostic bronchoscope biopsy procedure(including 7 patients underwent TBNA also). Among 41 patients, 3 underwent following cervical or axillary lymph node biopsy, 2 underwent parotid gland biopsy, 1 underwent gastroscopy; 3 underwent a CTguided transthoracic needle biopsy procedure to con rm the subtype classi cation or evaluate the neoplasm stage. A total of 20 patients were diagnosed with PPL. Among 41 patients, 18 cases were regarded as diffuse large B-cell lymphoma(DLBCL), 12 were extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue lymphoma(MALT), 5 were Hodgkin lymphoma(HL) ( Table 1). The predominant symptoms experienced by patients were coughing (36.6%), fever (17.1%), dyspnea (22%) and, while 48.8% patients suffered with symptom B. An interesting nding is high percentage of underlying chronic disease (chronic hepatitis B 12.2%, chronic hepatitis C 7.3% and Sjögren's syndrome 9.8%), which are summarized in Table 1. of SPL involved neck, stomach, liver, spleen, intestine and pelvic regions, et al. The SUV max of the pulmonary lesions ranged between 0 and 27.9, with 9.6 as the median value (Table 2).  The most common pathological diagnosis in our study is DLBCL, MALT and HL. Our study also showed high morbidity of chronic hepatitis B, chronic hepatitis C and Sjögren's syndrome in patients with pulmonary lymphoma, as 3 out of 4 Sjögren's syndrome patients were pathologically diagnosed as MALT.
Pulmonary lymphoma has speci c radiographic patterns and endobronchial changes. Rose described two distinct patterns of lymphomatous involvement: Type 1 occurs in patients with clinically apparent systemic lymphoma, whose common symptom is pneumonitis. The airway involvement is diffuse submucosal nodules in airway, while other radiographic ndings include parenchymal in ltrates, mediastinal lymphadenopathy. Some authors have suggested a preferential distribution at airway bifurcations[8]. Type 2 consists of a solitary mass involving the central airways without evidence of systemic lymphoma, while the symptoms are airway obstruction, wheezing or cough. Other radiographic ndings include atelectasis, lymph node enlargement in proximity [9].
In our study, nearly half patients are type 1, that the major CT abnormality was consolidation accounts for 24.4%, and GGO accounts for 24.4%, may involve parenchymal and lymphomatous in ltrates. (1)As Fig. 1a,b showed, for consolidation lesions, endobronchial patterns under bronchoscopy are usually mucosae asperity or edema, occasionally focal solitary mass. The air bronchogram sign and angiogram sign are more characteristic manifestations of this disease. Its pathological basis is that the neoplasm originated from pulmonary interstitium, while the pulmonary blood vessels still remain. The neoplasm grows along the bronchial wall, so the original bronchial remains within the lymphoma. It was previously noted that MALT are associated with air bronchogram in nearly 50% of cases [1]. While the hyperplasia of brous tissue can lead to traction bronchiectasis, that in consolidation lesions, bronchiectasis is another feature in pervious study [12]. (2)For GGO lesions, endobronchial patterns under bronchoscopy are always normal(except only one case found focal solitary mass under bronchoscopy), as Fig. 1c,d showed, while TBLB can usually make the histological diagnosis. Another half patients are type 2, that the major CT abnormality was mass accounts for 43.9%, and nodule accounts for 7.3%. (3)For mass lesions, with or without cavitation, endobronchial patterns under bronchoscopy are diffuse airway submucosal nodules or thickening, frequently focal solitary mass, as Fig. 1e,f showed. It is considered that cavitary lesions suggest a higher grade lymphoma, while consolidation suggests a lower grade lymphoma with reserved bronchial tree. (4)For nodule lesions, endobronchial patterns under bronchoscopy are bronchiole stenosis, as Fig. 1g,h,i showed, frequently mucosae asperity or edema. We found bronchial obstruction and associated atelectasis, ultimately atelectasis is usually suggestive of concomitant endobronchial disease. It seems the frequency of nodule lesions is much lower than previous studies mentioned above, because nodules are always not accessible by bronchoscopy. However, there is tendency for these CT or endobronchial patterns to overlap, and patients may present with several simultaneously.
Since 2008 World Health Organization classi cation, subtype classi cation of lymphoma is necessary for an accurate diagnosis, to predict disease prognosis, and to guide the treatment [13]. Our study excluded cases of suspected pathological diagnosis of lymphoma, but included 7 cases without subtype classi cation, that they needed following extrapulmonary biopsy. Previous studies suggested bronchoscopy may be of limited diagnostic value for pulmonary lymphoma diagnosis [14]. For example, Ferraro et al. reported that bronchoscopy was performed in 39 patients with PPL but con rmed the diagnosis in only seven patients (18%) [15]. The sensitivity of bronchial and transbronchial biopsies in detecting MALT lymphoma have been reported to be 30 and 88%, respectively in different researches [1,16].
However, for EBUS-TBNA, the diagnostic sensitivity seems better: the diagnostic sensitivity in lymphoma varied between 38% and 91% in the previous studies, according to diverse diagnostic criteria of the individual studies, proportion of new or prior lymphoma, prevalence of subtypes and implementing ancillary methods [17,18]. For example, involved 65 cases with nal diagnosis of lymphoma, Iqbal et al. indicated that the diagnostic sensitivity of EBUS-TBNA for patients with and without previously diagnosed lymphoma was 55% and 22%, respectively [18]. Other researches studies the use of new techniques as transbronchial needle forceps or ROSE, indicating the bene t that the lymphoma diagnosis undertaken by an expert group working closely with lymphoma pathologists and a lymphoma MDT [19,20]. Gonzalo et al. revealed the result of a meta-analysis, which involved fourteen studies with 425 patients. EBUS-TBNA reported an overall sensitivity of 66.2% and speci city of 99.3%. For a new diagnosis of lymphoma, thirteen studies including 243 participants reported sensitivity of 67.1% and speci city of 99.6%; for recurrence of lymphoma, eleven studies included 166 participants reported sensitivity of 77.8% and speci city of 99.5%. In the recurrence group, the use of ROSE, sample size and ow cytometry was found to increase the sensitivity of EBUS-TBNA, albeit a potential sources of heterogeneity [21].
Our study included de nitive diagnosis of lymphoma under bronchoscopy, that only 7 patients underwent TBNA meanwhile, probably because of the low sensitivity of EBUS-TBNA without ROSE. From my point of view, the use of bronchoscopy to diagnose lymphoma is easy to perform and minimally invasive with low frequency of procedure-related complication. Approximately one-third of patients can be given a de nitive diagnosis of lymphoma, with carefully patients selected according to CT features mentioned above(mass, GGO or consolidation), avoiding the need for further invasive procedures, such as CTguided transthoracic needle biopsy or thoracoscopy. Second, with high diagnostic accuracy in both benign and malignant diseases, bronchoscopy with or without EBUS-TBNA provides high differential diagnostic yield for lesions mimicking lymphoma such as lung cancer, infection and sarcoidosis [22]. Third, either a de nitive diagnosis or suspected lymphoma was provided currently for 85% of patients, which may be helpful for decisions with the management or further diagnostic procedures.

Conclusions
Pulmonary lymphoma presents distinct features in CT scan, which may indicate certain endobronchial patterns under bronchoscopy. The closer observation of CT images may facilitate the differential diagnosis and help in carefully patients selecting for bronchoscopy. This rst-line procedure may improve the early diagnosis and guide further individual clinical evaluation, aiding the early initiation of therapy strategies. Authors' contributions S-T Wang designed the study, had full access to all the data in this study, performed statistical analysis and wrote the draft. All authors read and approved the nal manuscript.

Funding
This research did not receive any speci c grant from funding agencies in the public, commercial, or notfor-pro t sectors.

Availability of data and materials
The datasets used for the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
This study was conducted in accordance with the amended Declaration of Helsinki. The ethical committee of Peking Union Medical College Hospital approved the protocol. Written informed consent was obtained from all subjects.

Consent for publication
Not applicable.