Pleural Invasion in Subsolid and Solid Lung Cancers: Predictive Features at CT and Their Clinical Significance

See also the article by Heidinger et al in this issue.

Brett M. Elicker, MD, is a clinical professor in the department of radiology and biomedical imaging at the University of California, San Francisco and chief of the cardiac and pulmonary imaging section. His clinical and research interests are in the areas of diffuse lung disease and lung cancer.

Introduction

CT and PET play a central role in determining the preoperative stage of patients suspected of having or known to have lung cancer. CT and PET are complementary in that they provide different information, CT delineating anatomic detail and PET delineating physiologic detail. Accurate staging is critical in determining the most appropriate treatment plans and in maximizing patient survival. While imaging has shown less than ideal correlation with the final pathologic staging, its primary role is to provide a noninvasive means of identifying lesions that may increase a patient’s stage. This has the potential to change management, either necessitating different surgical approaches or suggesting that radiation and/or chemotherapy are more appropriate than resection.

In patients whose preoperative staging suggests a lesion amenable to surgical resection, information provided by imaging also may contribute to the specific type of surgical procedure deemed appropriate: lobectomy, segmentectomy, or wedge resection. While lobectomy is considered the reference standard treatment for resectable lung cancer, many surgeons opt for sublobar resection in patients with smaller-diameter early stage lung cancers, as some evidence suggests no survival benefit of lobectomy over sublobar resection in these cases (1). On the other hand, it is important to note that there is some recent evidence suggesting that lobectomy is superior to sublobar resection, even for small early stage lung cancers (2). The choice of lobectomy or sublobar resection is particularly relevant for patients in whom sublobar resection conveys other advantages. These include patients with poor lung function and those who may require additional future surgical procedures. In the era of lung cancer screening and with the realization of a genetic predisposition for multiple lung cancers in certain patients, the potential need to treat multiple separate primary lung cancers must be considered in the selection of an appropriate surgical procedure.

The components of the preoperative T stage as determined with CT (size, invasion of adjacent structures, associated atelectasis, etc) have a particular impact on these surgical decisions. According to Detterbeck et al (3), pleural invasion is considered T2 (visceral pleural invasion) or T3 (parietal pleura and chest wall invasion). In general, pleural invasion has been associated with decreased survival and a greater risk for lymph node involvement. In one meta-analysis (4), patients with pleural invasion had a significantly decreased overall survival (hazard ratio = 1.555 to 2.447 depending on the depth of pleural invasion) compared with those who did not. Additionally, pleural invasion may suggest the possibility of a more aggressive tumor that has a greater propensity for regional or distant spread. In one study of 9297 surgically resected T1–2aN0M0 lung cancers, pleural invasion was associated with a greater need for more extensive lymph node sampling (5).

The ability of CT to predict pleural invasion in solid tumors has been investigated by several authors. Some of the features that have been associated with an increased risk of pleural invasion include the ratio of the length of pleural contact and total diameter of the tumor, a convex border of the tumor adjacent to the region of pleural contact (6), and the presence of pleural tags adjacent to a nodule (7). The CT features associated with pleural invasion in subsolid nodules are less well studied.

In the current issue of Radiology: Cardiothoracic Imaging, Heidinger et al (8) compared the CT features predictive of visceral pleural invasion in solid and subsolid nodules. As expected, the rate of pleural invasion was higher in solid tumors (24.6% of solid tumors vs 4.7% of subsolid tumors). In solid nodules, features that were associated with a higher risk of pleural invasion included the absolute length of pleural contact, the ratio of the length of contact and the diameter of the nodule, and more than one morphologic type of pleural tag. These morphologic tag types were previously defined by Hsu et al (7).

In subsolid nodules, features that were associated with a higher risk of pleural invasion in the article by Heidinger et al included the absolute length of pleural contact with the solid component of the nodule and the ratio of the length of contact of the solid component with the pleura and the overall diameter of the nodule. The odds of pleural invasion increased 1.27 times for each 1 mm of extra contact between the pleura and solid component of the tumor. Use of a threshold of ≥5 mm of contact between the pleura and the solid component of the tumor yielded a sensitivity of 100.0% and a specificity of 61.8%. A threshold of ≥12 mm yielded a lower sensitivity (60.0%) but a higher specificity (92.2%). Extension of the tumor through the fissure and infiltration of the extrapleural soft tissues surrounding the nodule were also associated with pleural invasion. The presence of either of these two later findings was 97.1% specific for pleural invasion.

To my knowledge, only two other studies have investigated the CT features of pleural invasion in mixed solid and ground-glass nodules. Ahn et al (9) showed that on multivariate analysis the following features were independently associated with pleural invasion in part-solid nodules: the presence of pleural contact, the presence of pleural thickening, solid portion comprising more than 50% of the nodule, and nodule size greater than 20 mm. Zhao et al (10), on the other hand, demonstrated that larger tumor size (≥2–3 cm) and nodules that indented, touched, or were within 1 cm of the pleura were associated with an increased risk of pleural invasion.

The study by Heidinger et al provides additional guidance to the radiologist and surgeon on the use of CT in predicting pleural invasion and suggesting when lobectomy may be preferable to sublobar resection. It also suggests situations in which a more extensive mediastinal lymph node sampling might be appropriate. These interventions should be considered when subsolid nodules demonstrate a larger solid component contacting the pleural surface and when there is either extension through the pleura or infiltration of adjacent tissues.

The need for these more aggressive interventions should always be weighed against the advantages of lung sparing surgery, and thus a multidisciplinary discussion incorporating input from multiple different subspecialties (surgery, radiology, oncology, and pulmonary medicine) is ideal for maximizing patient benefits. As part of this discussion, radiology plays a role in discussing several elements, including the risk of a given nodule representing malignancy, its potential TNM stage, the best approach for obtaining diagnostic tissue, and the most appropriate treatment plan. While a variety of other factors must be considered in this discussion, including other morbidities, pulmonary function test results, patient preferences, and so forth, these decisions are heavily based on the results of imaging tests. For this reason, radiologists must have an extensive knowledge of the utility of imaging in the preoperative staging of patients suspected of having lung cancer. Studies, such as the one by Heidinger et al, form the foundation of our understanding of this critical role of imaging.