Between August 2016 and December 2017, we enrolled 202 patients with stage I-III breast cancer who underwent preoperative 18F-FDG-PET-CT followed by primary surgery at Gangnam Severance Hospital, Yonsei College of Medicine, South Korea. In these patients, we successfully evaluated the SUVmax values and stromal TIL levels. To circumvent the effect of chemotherapy on SUVmax values and TIL levels, patients who underwent preoperative chemotherapy were excluded. To address more clear relationship between SUVmax and stromal TIL levels by reducing the partial-volume effect of PET35, we included only patients with tumors measuring ≥1 cm.
Clinical data on age at the time of surgery, HG, NG, tumor size, ER status, PR status, HER2 status, AR percentage, and Ki-67 L.I. were collected from the medical database. Tumors were classified according to the tumor–node–metastasis staging of the American Joint Committee on Cancer, 7th edition, and tumor grade was determined using the modified Scarf-Bloomer-Richardson grading system36. The study was approved by the Institutional Review Board (IRB) of Gangnam Severance Hospital (Local IRB number: 2020-0950-001), which waived the requirement of informed consent due to the retrospective study design. The study was performed in accordance with good clinical practice guidelines and the Declaration of Helsinki.
Assessment of TIL levels
TIL levels were measured as described in previous studies37,38. A pathologist (Y.J.C.) performed hematoxylin and eosin staining to review the histological features of treatment-naïve surgical specimens. Stromal TIL levels were evaluated according to the standardized methodology proposed by the international TIL Working Group11. The tumor area, defined by the presence of invasive tumor cells, was identified. All mononuclear cells including lymphocytes and plasma cells, but not polymorphonuclear leukocytes, were counted. The areas outside the tumor border, around the intraductal component, and around the normal lobules were excluded. Within the tumor border, areas showing crush artifacts and necrosis were also excluded. For each case, the average TIL level was measured using a representative section of the whole tumor, and the average level was reported as a percentage.
In this study, the cutoff value of high TIL level was set as 40%, which was used to analyze associations between TIL levels and pathological CR (pCR) or event-free survival in the NeoALTTO study39.
The procedure for 18F-FDG-PET-CT was the same as that previously reported20. After at least 8 hours of fasting, patients received an intravenous injection of 18F-FDG (0.14 MBq) in the arm contralateral to the primary tumor. Sixty minutes after injection of 18F-FDG, whole-body positron emission tomography scans were obtained using a Philips Allegro PET camera (Philips Medical Systems, Cleveland, OH, USA). During the scans, patients were placed in the supine position with their arms raised. The SUV was calculated by measuring the 18F-FDG uptake by the primary tumor in the region of interest using the following formula: SUV = (maximal radioactivity concentration in the region of interest)/(injected dose/patient’s weight (kg)). The SUV cutoff value of 4 was determined according to previous studies20,30.
Pearson’s correlation coefficient was calculated to measure the correlation between continuous TIL levels and SUV. According to tumor characteristics, mean TILs levels were compared using the Mann–Whitney U test or Kruskal–Wallis test. The Kolmogorov–Smirnov test was used to test the normal distribution of TIL data. The distributions of nonparametric variables were compared using the Mann–Whitney U test or Kruskal–Wallis test. The clinicopathological factors associated with high TIL levels (≥40%) were analyzed using logistic regression analysis. Variables that showed a significant difference (p<0.10) in univariate analysis were entered in multivariable analysis. To evaluate the ability of continuous SUVmax to predict TIL levels ≥40%, we determined the AUC using ROC curves. All statistical tests were two tailed, and p<0.05 was considered statistical significant. All statistical analyses were performed using SPSS, version 26.0 (SPSS, Chicago, IL, USA).
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.