Role of CT in differentiation between subtypes of lung cancer; is it possible?

Context and purpose: lung cancer is the second in the incidence rate and the first in death rate in the United States of America in 2017. Its treatment depends upon the tumor staging as well as the histological subtype of lung cancer. CT has been the modality of choice for screening as well as diagnosis of lung cancer; however, few studies tried to correlate different CT features of lung cancer to certain pathological subtypes. Our study aims to assess the CT characteristics of the subtypes of bronchogenic carcinoma. SQCC shows a higher incidence of central location compared with the rest of the lung cancers (significance level of 50%, p value of 0.5), internal cavitations (significance level of 94.9%, p value of less than 0.05) as well as more frequency of higher stage within the study population, ADC shows significant predilection to peripheral location compared with the rest of the lung cancers (significance level of 94.9%, p value of less than 0.05). There is an evident correlation between the MDCT diagnosis of bronchogenic carcinoma and that of histopathology/cytology. The most common types are SQCC and ADC subtypes. The SQCC type of bronchial carcinoma tends to be central with the internal cavitations are common while ADC tends to be peripheral and solid.


Background
According to the American Cancer Society (ACS), lung cancer is the second in the incidence rate and the first in death rate in the United States of America in 2017 [1]. In 2019, over 228,000 adults in the United States will have been diagnosed with lung cancer, and lung cancer constitutes approximately 13% of all new cancer diagnoses. It is estimated that over 76,000 men and over 66,000 women will die of lung cancer in 2019 [2].
In the absence of screening, most patients with lung cancer are not diagnosed until later stages, when the prognosis is poor. The most common symptoms are cough and dyspnea, but the mo3st specific symptom is hemoptysis. Digital clubbing, though rare, is highly predictive of lung cancer. Symptoms can be caused by the local tumor, intrathoracic spread, distant metastases, or paraneoplastic syndromes. Clinicians should suspect lung cancer in symptomatic patients with risk factors [3].
Radiologic manifestations of bronchogenic carcinoma include obstructive pneumonitis or atelectasis, lung nodule or mass, apical mass, cavitated mass, or nodule or mass associated with lymphadenopathy [4].
The initial study should be chest x-ray, it is readily available and inexpensive but if results are negative and suspicion remains, the clinician should obtain a computed tomography scan with contrast. Tissue samples should be obtained using the least invasive method possible [3].
CT of the chest is an important informative tool that helps in detailed imaging of the primary tumor and its anatomic relationship to other structures, and it provides information with respect to the size of mediastinal lymph nodes and the status of the pleural space. However, CT criteria for adenopathy are based on size alone and do not always accurately reflect the presence or absence of tumor metastases. CT can best be thought of as a technique that provides a roadmap for more accurate surgical staging [5].
The World Health Organization (WHO) classification applies to the surgically resected malignant tumors of the lung and pleura [1]. Primary carcinomas of the lung are traditionally classified as either small cell lung cancer (SCLC) or non-small cell lung cancer (NSCLC). NSCLC constitutes approximately 80% of all primary lung cancers with adenocarcinoma, squamous cell carcinoma (SCC), and large cell carcinoma constituting the major histological types [6,7].
As treatment depends upon the tumor staging as well as histological subtype of lung cancer [5] and since CT has been the modality of choice for screening as well as diagnosis of lung cancer, recently few studies tried to correlate different CT features of lung cancer to certain pathological subtypes.
Because these CT features have not been fully investigated, the purpose of our research was to compare the clinical pathology with different morphologic CT features.

Objectives
Assess the CT characteristics of the subtypes of bronchogenic carcinoma.

Methods
Data were retrospectively collected from high-resolution CT scans of 38 patients diagnosed as lung cancer from January to September 2018 in the Radio-diagnosis Department of Medical Research Institute. These patients underwent diagnostic pathological biopsy either CT in peripherally located lesions or bronchoscopic guided in central lesions.

CT protocol and image analysis
The CT scans were obtained on Siemens Emotion 16 multi-detector CT (MDCT) using volumetric High-spatial-frequency kernel algorithm with slice thickness of 1-1.25 mm, table speed for volumetric HRCT to enable the least cycles of breath-holds as possible, mean tube rotation of 0.75 s, collimation 1 mm, pitch 1.5, helical mode (volumetric HRCT), field of view (FOV) for small, medium, and large patients, kVp and mA per slice: 140 for each. For adequate multi-planar reconstruction, scans were performed to cover the root of the neck down to the level of the adrenal gland. Then, the images acquired were sent to a separate workstation (OsiriX V. 8.5) to be processed, manipulated, and reconstructed.
Multi-planar, as well as multiple intensity projection (MIP) reconstruction methods, had been done.
Using the blind technique, two radiologists with 12 and 15 years of experience fully revised and analyzed the thin-section CT scan findings including location (central or peripheral), size, shape, margin, involved lobe and the presence or absence of degeneration and cavitations, consolidation/atelectasis, air broncho-gram, nodal metastasis, vessel amputation, pleural/mediastinal infiltration, satellites and distant metastasis including suprarenal, liver and bone. Staging of the tumors was done following the eighth edition TNM stage classification for lung cancer [8].
The age of the study population ranged from 38 to 71 years with a mean of 55+/−9.2 years, no significant difference regarding the age was noted between the various pathological types of the study population with the mean age for squamous cell carcinoma patients was 55=/−10 while that for the adenocarcinoma patients was 59.5+/ −8.3 years.
The size of the lesions in the current study ranges from 1.6 to 25 c with mean size of 7.1+/−4.5 cm, squamous cell carcinoma lesions were larger than adenocarcinoma lesions with a mean size of 7.7+/−2.7 cm versus 5.5+/−2.3 cm respectively (Table 1).
Regarding the location of the lesion, the right upper lobe was the most encountered location with 15 lesions followed by the right lower lobe (9 lesions) then the left upper (7lesions ), left lower lobe (4 lesion), and the right middle lobe (one lesion) while one lesion was diffusely infiltrating the right lung (lung sarcoma).  SQCC shows a higher incidence of central location (64%) compared with the rest of the lung cancers (9.5%) (significance level of 50%, p value of 0.5) with the remainder central lesions were SCC and a single ADC (Table 1).
ADC shows significant predilection to a peripheral location (91.6%) compared with the rest of the lung cancers (9.5%) (significance level of 94.9%, p value of less than 0.05) ( Table 1).
Internal cavitations were noted in 47% of SQCC compared with 4% to the rest of lung cancers cases (significance level of 99.8%, p value of less than 0.002); in addition to these lesions, only one ADC was cavitary, the rest of the lesions were essentially solid; however, in the larger ones, there was internal cystic degeneration with peripheral air broncho-gram in the lesion proved to be invasive mucinous adenocarcinoma (Table 1).
In the current study, staging was done using The Eighth Edition Lung Cancer Stage Classification [9]. Mediastinal and pleural invasion was considered when the tumor tissue is seen within the mediastinal fat or within the pleural cavity, wide contact with either the  mediastinum or pleura (probable invasion) was not considered invasion in the statistical analyses nor staging.
Regarding the local staging of the lesions, SQCC has a higher incidence of T4 (70%) compared with the other types apart from SCC (the two included cases were T4); however, the difference was not significant ( Table 2). No significant statistical difference was noted regarding the pathological mediastinal nodes, pulmonary nor distant metastases between the different tumor types ( Table 2).

Discussion
Lung cancer kills more patients than any other malignancy in the world [10]. Histopathological analysis is the gold standard for diagnosing lung cancer and defines the cancer types. It is crucial to delineate lung malignancy from its morphologic mimic as specific treatment modalities (including surgical resection, chemotherapy, radiotherapy, and targeted therapy) can limit the progression of the disease and improve the survival outcomes of the patients [11].
SQCC was the most common subtype encountered in the current study followed by ADC accounting for 45% and 32% respectively, the distinction between them is critical for selecting the optimal treatment [12].
In the literature, SQCC is the most common subtype of bronchial carcinoma; however, the incidence of ADC is increasing, also 11 studies reviewing earlier diagnoses  increased ADC by 30% overall. This may be due to changes in the diagnostic criteria and changes in bronchial carcinoma classifications [13]. Moreover, Zhang et al. [14] found in their study based on data of 658 primary lung cancer patients that ADC was the most common subtype followed by SQCC then SCC, the proportion of ADC increased from 25.93 (1995-1997) to 56.36% (2013-2015). The number of SQCC cases increased from 710 (1995)(1996)(1997) to 3050 (2013-2015); however, the proportion decreased from 49.1 to 26.34%. Among the current study population, 13 lesions were central in location (34.2%) from which 11 were SQCC, one ADC, and one SCC.
In the literature, the incidence of central of SQCC is well known in many studies including Mizushima et al. [15] who found in their study over 235 squamous cell carcinomas that 129 was peripheral and 106 were central and William Krimsky et al. [16] found a total of 56 patients were diagnosed with SCC. Of these, 55% (n = 31) had peripheral and 45% (n = 25) had central SCC.
Mismatching results were documented in the study by Zhe et al. [17] were in 302 consecutive patients included 99 patients were ADC, 95 patients with SQCC, and 108 patients with SCLC with ADC were more aggressive.
Regarding the SCC, Dongiun et al. [18] in his study on 142 lesions, 121 from 142 lesions to be central from which 112 lesions invade the mediastinum.
Internal cavitations were noted in 9 cases of the current study population including 8 SQCC and a single ADC. Matching results were described by Chaudhuri et al. [19] who found in their study on 100 cavitary lesions that 82 were SQCC, 11 undifferentiated carcinomas of large polygonal-cell type, and seven ADC.
In the 2011 classification [20], ADC was divided into pre-invasive types including atypical adenomatous hyperplasia and ADC in situ, minimally invasive, and invasive types. Pre-invasive types appear on CT commonly as ground-glass attenuation, minimally invasive and invasive types commonly appear with the solid invasive component that varies in size (histologically in noninvasive it is less than 5 mm and in invasive type, it is more than 5 mm) [21].
Other CT features of ADC were described in the literature including cavitations, calcification, consolidations, and scar-like appearance [22]. ADC in the current study was noted in 11 peripheral lesions and one central lesion, all these lesions were invasive type with two of them showing anaplastic transformation. Radiologically, one lesion was cavitary, 11 were solid and there was a rim of ground-glass attenuation noted around 5 of the lesions reflecting bronchoalveolar components; the solid lesions included 6 with speculated appearance, one was stellate like three shows lobulated margins, and one with peripheral air bronchogram proven pathologically to be invasive mucinous adenocarcinoma (formally was described as bronchoalveolar carcinoma (BAC) ) [23].
The current study included two carcinoid tumors, both were large lesions with lobar endobronchial extension, Our findings match with the data in the literature as most of the larger studies [24,25] of pulmonary carcinoid tumors have found that 85% of lesions arise within the central airways as endobronchial masses, aside from an article by Magid et al. [26] in 1989 and Quinn et al. in 2011 [27].
In the current study, two large cell carcinomas were included, both were solid and peripheral in location with no evident pleural involvement. In the literature, Oshiro et al. found in his study on 38 clear cell carcinoma lesions that 32 lesions were peripheral and 6 lesions were central in location, the mean size of the lesions was 3.2 cm with lobulated outlines and homogeneous solid nature apart from cystic necrosis in larger tumors [28]. A single sarcoma was included in the current study, and it was a large solid lesion inseparable from the pleura and mediastinum and occupying most of the right hemithorax. Limited studies noted in the literature on the lung sarcoma with Duran et al. [29] found in his study on 7 lesions that lung sarcomas are large and homogeneous.

Conclusion
There is an evident correlation between the MDCT diagnosis of bronchogenic carcinoma and that of histopathology/cytology. The most common types are SQCC and ADC subtypes. The SQCC type of bronchial carcinoma tends to be central with the internal cavitations are common while ADC tends to be peripheral and solid.