Clinical characteristics and work-up of small to intermediate-sized pulmonary nodules in a Chinese dedicated cancer hospital

Study population

The study was conducted at the Tianjin Medical University Cancer Institute and Hospital (TJMUCH), one of the most important specialized cancer centers in Northern China, which treats thousands of lung cancer patients each year. A consecutive search was performed using the PACS database on patients at TJMUCH receiving chest CT scans in 2013. The following inclusion criteria were applied: 1) all pulmonary nodules were newly detected in 2013 and 2) nodule diameters ranged from 4 to 25 mm. The final diagnosis of nodule nature (malignant or benign) was confirmed based on histological results or follow-up. Additionally, nodules were considered benign when they had disappeared, reduced in size by at least 30% or remained stable for at least 2 years on CT follow-up¹¹. Exclusion criteria were as follows: 1) patients with suspected metastatic disease based on chest CT or histopathological findings and 2) nodules with unconfirmed results (for instance, no histology and inadequate follow-up period). Data on patients, nodules and clinical characteristics were collected from the hospital information system. Basic patient data included: 1) gender; 2) age; 3) smoking history; 4) cancer history; 5) family cancer history. Information on nodule characteristics included: 1) location (upper, middle or lower lobe); 2) size (average value of maximum long diameter and short diameter measured in the lung window); 3) shape (round, oval, irregular); 4) margin (smooth, lobulated, spiculated); 5) nodule type (non-solid, part-solid, solid); 6) density (attenuation value of the largest region of interest). Details of the assessment are described in Supplementary Figure S1–S5. Clinical characteristics included: 1) clinical diagnostic examination (CT, PET-CT, and biopsy); 2) surgery characteristics; 3) pathology results; (4) gene mutation information.

CT protocol

Chest CT examinations were conducted using a Discovery CT750 HD (GE Medical Systems, Milwaukee, WI), Lightspeed 16 (GE Medical Systems) or Somatom Sensation 64 (Siemens Medical Solutions, Forchheim, Germany) CT system. The CT examination encompassed regions ranging from the pulmonary apex level to below the diaphragm. For scanning, tube voltage of 120 kVp with automatic tube current modulation was applied. For the GE CT system, reconstructed slice thickness and pitch were 1.25 mm and 0.984, respectively. For the Siemens CT system, reconstructed slice thickness and pitch were 1.5 mm and 0.95, respectively. The first CT examination was generally a non-contrast-enhanced CT. For further nodule evaluation, contrast-enhanced CT usually was performed at a later time. For contrast-enhanced CT, the iodine contrast agent Visipaque (Iodixanol, 270 mg/mL) was administered at a concentration of 1.5 mL/kg and injection rate of 2.5 mL/s. Contrast agents were administered intravenously through the upper extremity. Scanning was performed 70 s after the start of injection.

Statistical analysis

Data are presented as numbers (%) for categorical variables. Pearson’s Chi-square test was used to examin differences in categorical variables between benign and malignant nodules. Normally distributed continuous data are presented as mean [standard deviation (SD)] and compared using Student’s t-test. Non-normally distributed continuous variables are presented as median [inter-quartile range (IQR)] and compared using a non-parametric test. P-values < 0.05 were indicative of statistical significance. All statistical analyses were performed using SPSS software version 20.0 (IBM, New York, US).

Basic characteristics

In total, 491 patients with 609 primary pulmonary nodules fulfilled the inclusion criteria. Owing to lack of diagnostic confirmation, 295 nodules in 192 patients were excluded, leading to the final inclusion of 299 patients with 314 nodules for analysis. Out of the 299 patients, 206 (68.9%) were diagnosed with lung cancer (Figure 1). Within this patient group, 75 (72.1%) men and 16 (15.7%) women were smokers at the time of the study. Among 65 nodules subjected to EGFR mutation analysis, 13 (20%) were identified as EGFR19, 15 (23.1%) as EGFR21, and 37 (56.9%) as EGFR wild-type. Patient characteristics are presented in Table 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Flowchart of study population inclusion.

Nodule evaluation

In total, 212 of the 314 small to intermediate-sized nodules were malignant (67.5%). CT characteristics are presented in Table 2. The majority of nodules were solid (n = 262, 83.4%), among which 168 (64.1%) were malignant, while 34 out of the 36 (94.4%) partly solid nodules were malignant. The maximum diameter of malignant nodules was 18.0 mm (median, IQR: 14.0–22.0 mm), which was significantly larger than that of benign nodules (median 12.5 mm, IQR: 9.0–18.0 mm). Nodule margins differed significantly between malignant and benign nodules. Malignant nodules often were spiculated [n = 153 (72.2%)] while 46 (45.1%) of the benign nodules were smooth (P < 0.001). Nodule densities of benign and malignant nodules detected using non-contrast enhanced CT were similar. However, malignant nodules detected with contrast-enhanced CT had significantly higher density [67.0 HU (IQR: 51.0–81.0)] than benign nodules [57.5 HU (IQR: 34.5–77.0, P = 0.015)] (Figure 2 and Figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Patient example 1. Male, 64 years, a solid spiculated nodule (13 mm * 11 mm) located on the right upper lobe with the pleural indentation sign. The mean density of the nodule was 16 Hounsfield Units on CT without iodine contrast, and uneven enhancement 97 Hounsfield Units after contrast (A and B: non-contrast-enhanced CT. C: contrast-enhanced CT). Histology: Adenocarcinoma (stage I).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Patient example 2. Female, 58 years, a solid smooth nodule (18 mm * 14 mm) located on the right upper lobe and close to the vessel. The mean density of the nodule was 41 Hounsfield Units on CT without iodine contrast, and uniform enhancement 52 Hounsfield Units after contrast (A and B: non-contrast-enhanced CT. C: contrast-enhanced CT). Histology: Hamartoma.

Management and histopathological diagnosis of pulmonary nodules

Histopathological diagnosis was obtained for 268 out of 314 (85.4%) nodules. In 232 (73.9%) cases, the nature of nodules was confirmed by surgery and 36 (11.5%) cases by biopsy. Table 3 presents an overview of nodule work-up. Among the 102 benign nodules, 66 (64.7%) were confirmed by surgery, 6 (5.9%) by biopsy, and 30 (29.4%) by stable size or resolution at CT follow-up. Nodules confirmed as benign by surgery were significantly larger than those determined as benign after CT follow-up (14.0 mm vs. 9.0 mm, respectively, P = 0.001).

Before surgery, the majority of patients only received a contrast-enhanced CT scan after an initial non-contrast CT scan for nodule staging in both malignant [n = 140 (84.3%)] and benign [n = 54 (81.8%)] cases. Thirty-two of the patients with confirmed malignancy (19.3%) underwent a PET-CT scan before surgery, 2 (1.2%) underwent biopsy before surgery, and 12 (7.2%) underwent surgery immediately after the initial non-contrast CT scan. Among the resected lung cancers, 128 (77.1%) were diagnosed at stage I, 8 (4.8%) at stage II, 25(15.1%) at stage III, and 5 (3.0%) at stage IV (2 cases of pleural metastasis, 2 bone metastasis, 1 brain metastasis).

In total, 196 nodules were histologically confirmed as malignant. The majority of lung cancers were adenocarcinomas [n = 163 (83.2%)] and squamous cell carcinomas [n = 18 (9.2%)]. Among the 72 resected benign nodules, the most common histological diagnosis was hamartoma [n = 20 (27.8%)]. Histopathological results are shown in Table 4.

Patient characteristics

In our cohort, 67.5% of the clinically detected nodules were malignant. This lung cancer rate is obviously higher than that reported from community-based lung cancer screening using low-dose CT in China (malignancy rate between 0.6% and 1.5%)^3,12,13. Moreover, relative to studies on other clinical cohorts, the percentage of malignant nodules in our analysis was considerably high⁹. This could be attributable to our hospital being a dedicated cancer center to which most patients are referred because of suspicious nodules. In fact, 192 patients with 295 nodules were excluded owing to unconfirmed diagnosis (Figure 1). We lost contact with patients once they left our hospital, which presents a major problem in the long-term management of patients with pulmonary nodules in China.

Mean age at lung cancer diagnosis was similar to that reported in an earlier Chinese multi-institutional registry (CMIR) study (59.5±10.2 vs. 58.9±10.7)³. Smoking is the main cause of lung cancer, accounting for 80% of the male lung cancer burden and at least 50% of the female lung cancer burden worldwide¹⁴. However, in China, more than 30% of patients with lung cancer are non-smokers^15–17. In our cohort, the proportion of smokers among female patients (15.7%) was substantially lower than that among male patients (72.1%). Several studies suggest second-hand smoking and gene mutations are risk factors for lung cancer in female non-smokers^18–20. The EGFR mutation rate in our study was > 43%, which was significantly higher than that recorded in the Western population (approximately 7%)¹⁰.

CT evaluation of pulmonary nodules

Radiologists generally estimate lung cancer probability on the basis of nodule CT characteristics. In our study, partly solid nodules showed a high probability of malignancy (94.4%), consistent with data from earlier screening studies^21–23. For solid nodules, pre-surgical diagnosis was usually based on contrast-enhanced CT. Malignant nodules generally showed a large density difference between contrast-enhanced and non-enhanced CT, which was negligible or smaller for benign nodules. In our patients, contrast enhancement on CT was one of the most important differentiators of malignant and benign nodules before surgery.

Nodule margin is another important feature in determining nodule malignancy risk. Malignant nodules tend to be spiculated and lobulated while benign nodules are more likely to be smooth. In our cohort, 5.0% of malignant nodules were smooth, 22.6% were lobulated and 72.2% spiculated (P < 0.001, compared to margins of benign nodules). Similar findings have been reported previously^24–27.

Nodule diagnosis and management

One of the most important objectives of nodule work-up is to diagnose malignant nodules as rapidly and accurately as possible while minimizing the number of surgically removed benign nodules and late-stage lung cancers. Among the 232 resected nodules, 166 (71.6%) were malignant and 66 (28.4%) were benign. Only 19.3% lung cancer patients in our study underwent PET-CT prior to surgery, which is significantly lower than previous reports (55.0%)⁹. Instead, multiple body part CT and MRI scans were often performed for staging. Thus contrast-enhanced CT was the main imaging examination technique employed prior to surgery. This is consistent with work-up guidelines in China based on expert opinions, the Diagnosis and Treatment for Lung Nodules (2016 Version)²¹, which provides a choice of contrast-enhanced CT, PET-CT, and biopsy. Among the benign nodules, only 30 (29.4%) were examined using CT follow-up in our study. The percentage of resected benign nodules was high, since 59 (57.8%) of all benign nodules were considered high risk according to the China National Guideline of Classification²¹. Furthermore, the diameters of resected benign nodules were generally larger than 10 mm. Considering the relatively high resection rates of benign nodules and late-stage lung cancers (stage III or IV) in our study, it appears essential to refine the diagnostic process by using PET-CT to avoid unnecessary surgery. At the same time, excessive application of PET-CT will increase the financial burden for patients. To reach a compromise, more clinical research and expert consensus are required for optimization of future guidelines for China.

Our study population may reflect a high degree of selection bias, since our patients were sourced from a dedicated cancer hospital, with most being referred due to suspicion of malignant nodules. In addition, we only included patients with available pathology results or more than 2 years of follow-up. Several patients with small nodules that were considered benign did not receive imaging follow-up in our center and were therefore excluded. These factors may explain the relatively high rate of malignancy in our study population relative to other studies. Our data did not include patient information on ‘second-hand smoking’ and therefore, the effect of passive smoking was not estimated. Moreover, the current report is based on retrospective analysis of patient data from a single center and a multicenter prospective study is lacking.