*2.1. Patient Population*

This retrospective study was conducted in accordance with the Declaration of Helsinki and the protocol was approved by the local Institutional Review Board and Ethics Committee at our hospital (IRB109-235-B). Due to the retrospective nature of this study, the requirement of informed consent for this study was waived. We retrospectively enrolled patients with a new diagnosis of NSCLC from January 2010 to September 2019. The diagnoses of NSCLC in all study patients were established using histopathology. Serial examinations were performed in all study participants for lung cancer staging and treatment planning at the time of the initial diagnosis. The examinations included contrast-enhanced computed tomography (CT) of the chest to the upper abdomen, 18F-FDG PET/CT, gadoliniumenhanced MRI of the brain, and pulmonary function tests. For lesions in the images that were indicative of malignancy, image-guided biopsies were collected whenever possible. If biopsies were not feasible or if the biopsy result was negative, the patient was closely monitored with imaging. Patients were re-staging according to the eighth edition of the AJCC staging manual [13]. We only included patients with a clinically positive regional

nodal metastatic disease (patients with an N1 to N3 classification were included) and without evidence of distant metastasis at the time of the initial diagnosis. Patients' daily living performance at the initial diagnosis was assessed using the Eastern Cooperative Oncology Group scale (ECOG) [24]. Patients included in this study received curative surgery (resection of the primary tumor and mediastinal lymph node dissection) with or without neoadjuvant CCRT, definitive CCRT, or definitive radiotherapy as the initial treatment. The radiotherapy dose was 2 Gy/fraction daily up to a targeted dose of 60 to 66 Gy. A cisplatin-based chemotherapeutic regimen was administered if CCRT was chosen as the initial treatment [25,26]. Patients receiving only systemic chemotherapy or target therapy were excluded. The choice of first-line treatment was based on the decision of the attending physician. The findings of all examinations in each individual and the pre-treatment staging were discussed and determined at a multidisciplinary lung cancer conference convened by our thoracic oncology research team.

#### *2.2. Imaging Protocol and Analysis of 18F-FDG PET Scan*

All study participants fasted for at least 4 h before 18F-FDG injection (400 MBq) and had blood glucose levels no greater than 200 mg/dL. The 18F-FDG PET/CT scans were performed 45 to 60 min after radiotracer administration using a GE Discovery ST scanner (GE Healthcare, Milwaukee, WI, USA). The PET/CT system was equipped with a PET unit containing 10,080 bismuth germanate crystals in 24 rings and a 16-detector row transmission CT unit. CT scans were performed first for attenuation correction without administration of contrast medium. The voltage and current of the tube were 120 kV and 120 mA, respectively. The pitch of the CT was 1.75 and sampling of CT images was done in the helical mode with a helical thickness of 3.75 mm. Immediately after the transmission CT, PET images were acquired from the midthigh to the vertex in a static 3-dimensional mode. The scanning time was three min for each table position (15 cm for each table position, with a 3 cm overlap for every contiguous frame). PET images were reconstructed with an ordered-subset expectation maximization algorithm (2 iterations, 21 subsets, and a 2.14 mm full width at half maximum Gaussian post-filter). The imaging matrix size, pixel size, and slice thickness for the reconstructed PET images were 128 × 128, 5.47 × 5.47, and 3.27 mm, respectively.

The platform used for display and semiquantitative analysis of 18F-FDG PET/CT images was a PMOD 4.0 system (PMOD Technologies Ltd., Zurich, Switzerland). An experienced nuclear medicine physician interpreted the 18F-FDG PET images. For image quantification, the experienced nuclear medicine physician identified and placed the volume-of-interest (VOI) on the primary tumor and the regional metastatic nodes on the 18F-FDG PET/CT image. The VOIs were placed and segmented separately for the primary tumor and the metastatic nodes. The SUV of 18F-FDG PET was calculated and normalized to each patient's body weight as follows:

$$\text{SUV} = \frac{(\text{decay} - \text{corrected activity} \ (\text{kBq}) \ \text{per milliliter of tissue volume})}{(\text{injected 18F} - \text{FDG activity} \ (\text{kBq}) / \text{body weight in g})}$$

The 18F-FDG lesions were segmented using a 41% threshold of the maximum standard uptake value method [27]. The segmented volumes were used to define the MTV. The PMOD 4.0 software automatically generated the mean SUV within the MTV. The TLG was then calculated as TLV = mean SUV × MTV. The VOI and segmentation results were confirmed by another expert nuclear medicine physician. We recorded the SUVmax and TLG values of the primary tumors (described as primary tumor SUVmax and primary tumor TLG) and the regional metastatic nodes (nodal SUVmax and nodal TLG). Furthermore, we calculated the nodal to primary tumor SUVmax ratio (NTSUVR), the nodal to primary tumor TLG ratio (NTTLGR), the product of the primary tumor and nodal SUVmax (TNSUVproduct), and the sum of the primary tumor and nodal TLG (total TLG) based on the following formulas:

NTSUVR = (nodal SUVmax) (primary tumor SUVmax) NTTLGR = (nodal TLG) (primary tumor TLG) TNSUVproduct = (primary tumor SUVmax) × (nodal SUVmax) total TLG = primary tumor TLG + nodal TLG The procedure used for image feature extraction is outlined in Figure 1.

**Figure 1.** The method used for feature extraction from 18F-FDG PET. VOI, volume-of-interest; SUV, standardized uptake value; TLG, total lesion glycolysis; NTSUVR, nodal to primary tumor SUVmax ratio; NTTLGR, nodal to primary tumor TLG ratio; TNSUVproduct, product of primary tumor and nodal SUVmax.

#### *2.3. Follow-Up of Study Participants*

After diagnosis, we followed patients with weekly outpatient clinic visits during treatment, at 3-month intervals after initial curative treatment, at 6-month intervals for 2 years, and annually thereafter. When signs of disease recurrence or progression emerged, contrast-enhanced CT, gadolinium-enhanced MRI of the brain, or 18F-FDG PET/CT were performed. Biopsies were taken for suspicious lesions whenever possible. New bloody effusion or positive fluid cytology was considered as recurrence or disease progression.
