Relationship Between Pulmonary Hamartoma and Bronchus: Multislice Spiral Computed Tomography Findings

Objective The aim of the present study is to investigate the relationship between pulmonary hamartomas (PHs) and bronchi on multislice spiral computed tomography (MSCT) images. Methods The MSCT scans of 218 PHs from 216 pathologically conrmed patients were reviewed. The PHs were divided into two groups, namely, the central endobronchial and intraparenchymal groups, in accordance with location. Multiplanar reconstruction was used to demonstrate PH–bronchus relationship patterns. The PH–bronchus relationships in the intraparenchymal group were classied into ve patterns: type I, the bronchus was cut off by the tumor; type II, the bronchus was contained within the tumor (air bronchogram sign); type III, the bronchus ran at the tumor periphery or was compressed by the tumor; and type IV, no tumor–bronchus relationship was observed.


Introduction
Pulmonary hamartomas (PHs) are the most common benign tumors; they account for approximately 6-8% of all solitary pulmonary nodules [1]. They are clinically divided into two types in accordance with location: intraparenchymal and endobronchial, with the latter being more frequent than the former [2]. If PHs are diagnosed through imaging, surveillance, rather than resection, is recommended given the absence of the evidence of malignant transformation during follow-up. Resection is considered if a lesion is clinically present or when a malignancy cannot be excluded [3]. Therefore, the imaging diagnosis of the most suggestive nding of PHs [1,4]. However, approximately 50% of PHs show neither fat nor calci cation on CT, complicating the differentiation of PHs from lung cancer or other benign nodules [1].
We hope to provide further value for the diagnosis or exclusion of PHs.

Study Population
This retrospective study was approved by the institutional Ethics Committee, and the requirement for informed consent was waived. We collected 218 PHs from 216 cases (2 of the cases had 2 PHs each) that had been histologically proven through surgery (213) or bronchoscopy (5) between May 2008 to April 2020. The patients (110 males and 106 females) ranged in age from 28 years to 76 years (average age of 56.4 years).

Evaluation of CT features
The locations of the PHs were recorded, and the PHs were divided into two groups, namely, the central endobronchial and intraparenchymal groups, in accordance with their location. Multiplanar reconstruction was used to demonstrate the patterns of PH-bronchus relationships.
The PH-bronchus relationships in the intraparenchymal group were classi ed into ve patterns: type I, the bronchus was cut off by the tumor; type II, the bronchus was contained within the tumor (air bronchogram sign); type III, the bronchus ran at the tumor periphery or was compressed by the tumor; and type IV, no tumor-bronchus relationship was observed.
The internal fat and internal calci cation of PHs in the central endobronchial group and in types I and II patterns in the intraparenchymal group were evaluated.
Two radiologists (with 10 and 12 years of experience in thoracic radiology) retrospectively reviewed the MSCT scans from PACS independently. Consensus was reached after mutual discussion in case of disagreement.

Results
Nine (4.1%) PHs were assigned to the central endobronchial group, and 209 (95.9%) PHs were assigned to the intraparenchymal group. For the endobronchial group, 1 (11.1%) PH was located in the trachea with the partial stenosis of the trachea (Fig. 1), whereas the remaining 8 (88.9%) PHs were located in the lobar (4 PHs) or segmental (4 PHs) bronchus with the complete occlusion of the corresponding bronchus. For 8 PHs in the lobar or segmental bronchus, 2 PHs were in the right upper lobe bronchus, 1 PH was in the right middle lobe bronchus, 1 PH was in the left upper lobe bronchus, 1 PH was in the posterior segment of the right upper lobe bronchus, 2 PHs were in the lingual segment of the left upper lobe bronchus, and 1 PH was in the dorsal segment of the left lower lobe bronchus. Two PHs contained fat and calci cation (Fig. 1), 3 PHs contained fat only, 2 PHs contained calci cation only, and 2 PHs contained no calci cation or fat.
For the intraparenchymal group, 46 (22.0%) PHs were in the upper lobe of the right lung, 20 (9.6%) PHs were in the middle lobe of the right lung, 53 (25.4%) PHs were in the lower lobe of the right lung, 47 (22.5%) PHs were in the upper lobe of the left lung, 37 (17.7%) PHs were in the lower lobe of the left lung, and 6 (2.9%) PHs grew across lobes (3 PHs spanned the middle and lower lobes of the right lung and 3 PHs spanned the upper lobe and lower lobes of the left lung). A total of 114 (54.5%) PHs were peripherally located at distances of less than 5 mm from the pleura, and 76 (36.4%) PHs were attached to the pleura. Among the patterns of intraparenchymal PH-bronchus relationships, type IV (147, 70.3%) was most common pattern, followed by type III (54, 25.8%) and type I (8, 3.8%) (Fig. 2)

Discussion
This was the rst large-sample study on the visualization of the PH-bronchus relationship on MSCT. We found that 88.9% (8/9) of central endobronchial PHs obstructed the bronchus, whereas only 3.8% (8/209) of intraparenchymal PHs obstructed the bronchus. No air bronchogram sign was observed.
Previous studies found that the sign wherein the bronchus is cut off by a lesion (type I) is signi cantly more common in lung cancers than in benign lesions [5,6,11]. Qiang et al. found that the type 1 pattern is shown by 58.5% of malignant nodules and by only 16.0% of benign nodules [6]. Choi et al. reported that type I is most common pattern in squamous cell carcinoma; it is presented by 45.8% of squamous cell carcinomas and is especially common in central squamous cell carcinomas (55%) [11]. Bronchus amputation in lung cancer can be ascribed to two reasons. One is the destructive power of the tumor. Tumor cells proliferate and invade the surrounding lung parenchyma, destroying the lesion area and adjacent bronchi [6]. The other reason is that lung cancer arises mainly from the bronchial mucosal epithelium; therefore, the bronchus is easily cut off in the early stage [14]. Amputated bronchi also occur in in ammatory nodules, such as granulomatous in ammation or common in ammatory lesions, due to the lling of bronchi by mucinous or granulomatous tissue, the in ammation of bronchi, or the destruction of bronchi by the lesion [5,6]. In this study, we found that the type I pattern was common in central endobronchial PHs (8, 88.9%) but was rare in intraparenchymal PHs (8, 3.8%). Central endobronchial PHs originate in the bronchi, leading to bronchial stenosis, peripheral in ammation, and the lling of bronchi by mucus. These phenomena lead to the type I sign. Only 1 central endobronchial PH in this study did not cut off bronchi because it was located in the trachea, which is wide, and the lesion itself was small. Moreover, no sign of infection was observed. We hypothesized that the bronchus might have been cut off by the intraparenchymal PH because of its origin in the bronchus. Speci cally, as the tumor grew expansively, it amputated the small bronchus. Central endobronchial PHs should be distinguished from malignant tumors, especially squamous cell carcinomas. Other relatively rare diseases include infectious lesions and foreign bodies. Among the 9 central endobronchial PHs in this study, only 5 contained fat (2 of which were accompanied by calci cation). This characteristic is helpful for the diagnosis of PHs on MSCT. Given that the type I sign was rarely observed for intraparenchymal PHs, we concluded that the presence of this sign in a solitary intraparenchymal lesion should preclude the rst diagnosis of PHs. However, the presence of fat and/or popcorn-like calci cation on CT is the suggestive nding of PHs. In this study, for type 1, fat was found in 2 intraparenchymal PHs and popcornlike calci cation was found in 3 intraparenchymal PHs. These characteristics are helpful for the diagnosis of PHs.
The air bronchogram sign is an important radiologic sign. Initial studies indicated that the presence of this sign is suggestive of pneumonia or benign nodules, but malignant tumors were also later found to be common [5][6][7][8][9][10][11][12][13]. The incidence of the air bronchogram sign in malignant solitary pulmonary nodules (28.7-81.5%) is usually higher than that in benign nodules (5-33.3%) [7][8][9][10]. In lung cancer, the air bronchogram sign is more common in adenocarcinoma, and different types of air bronchogram signs (normal, tortuous, and ectatic) are associated with the cell types and malignancy of adenocarcinoma [7,8,[10][11][12][13]. In the early stage of adenocarcinomas, the growth pattern of tumors is mainly lepidic, tumor cells spread along the alveolar wall and the interalveolar septum, and the bronchi are surrounded but not destroyed, leading to an air bronchogram sign [6,15]. Furthermore, with the retraction of tumoral brosis, the bronchi within the tumor are not only unobstructed and uncompressed but remain patent and even irregularly dilated [16,17]. As their malignant degree increases, tumors may encroach on the walls of bronchi, causing the bronchi to become irregular, narrow, or truncate. In addition, the air bronchogram sign can be seen in in ammatory nodules, such as nodules associated with ordinary in ammation, cryptococcus infection, and tuberculosis [6-9, 10, 18, 19]. Takanashi et al. reported that the air bronchogram sign is more frequently found in pneumonia (40%) than other lesions [8]. An air bronchogram sign is present in 63% of in ammatory pseudotumors in the ndings of Qiang et al. [6]. Given that the early formation of in ammatory nodules is related to the exudation and consolidation of the alveolar space, bronchi can exist without being destroyed or lled, and some in ammatory nodules can produce contractile force due to the proliferation of brous tissue in the chronic stage; this effect causes bronchiectasis. Bronchi may be natural or dilated, narrowed, or amputated when lled with granulation tissue, mucus, or bleeding [20]. The present study included 209 intraparenchymal PHs, and none presented the air bronchogram sign likely because of the expansive, rather than invasive growth of PHs. Thus, changes to the bronchi were pushing rather than enveloping. Therefore, we concluded that PH should not be diagnosed if the air bronchogram sign is present in a solitary lesion.
Benign and malignant nodules can present the sign of bronchi run at the tumor periphery or are compressed by the tumor; compressed bronchi with walls that remain soft and intact are more common in benign tumors, whereas bronchi with rigid and thickened walls are more common in malignant tumors [6]. Given that PHs cannot invade the parenchyma, the bronchi around the nodule are usually patent or compressed, whereas in malignant nodules, tumor cells always in ltrate around the bronchial wall, resulting in wall rigidity and thickening [6]. In this study, 54 (25.8%) intraparenchymal PHs presented bronchi that ran at the tumor periphery or were compressed by the tumor. None of these PHs had rigid and/or thickened bronchi. In addition, the majority (147, 70.3%) of intraparenchymal PHs in this study lacked a tumor-bronchus relationship likely because most of the intraparenchymal PHs were located in the periphery of the lung, and the peripheral bronchi were small and di cult to observe on CT. In this study, 114 (54.5%) of the intraparenchymal PHs were located at distances of less than 5 mm from the pleura and 76 (36.4%) of the intraparenchymal PHs were attached to the pleura.

Conclusion
This study systematically analyzed the PH-bronchus relationship on MSCT with a large sample. Central intrabronchial PHs were prone to cutting off bronchi because of their location within the bronchi. Intraparenchymal PHs, which were mostly located around the periphery of the lung, were mostly unrelated to bronchi. Only a few PHs cut off bronchi, and no air bronchogram sign was observed. Endobronchial PH in a 62-year-old man. The PH was located in the trachea with the partial stenosis of the trachea (arrow). The PH was lobulated with a smooth border and contained fat and calci cation