Influences of intratumoral heterogeneity on assessment of tumor microenvironment in esophageal squamous cell carcinoma

Abstract

The intratumoral heterogeneity (ITH) of the tumor microenvironment (TME) has yet to be addressed in esophageal squamous cell carcinoma (ESCC). Here, we studied the ITH of CD8 and PD‐L1 status in ESCC, and examined the potential of the tumor surface for representing the TME. In total, 67 surgically resected clinical Stage II ESCC specimens were analyzed. The CD8‐cell density, PD‐L1 tumor proportion score (TPS), and combined positive score (CPS) were calculated in three (superficial, middle, and deep) areas of each specimen. ITH was quantified by distance‐standardized coefficient variations of the three values. The CD8 and PD‐L1 status of each area was dichotomized and tumor‐surface capabilities for predicting the entire tumor status were estimated. Variables were compared according to the presence of neoadjuvant chemotherapy (NAC). The ITH, especially PD‐L1 heterogeneity, differed markedly among specimens. The concordance rates of CD8 and PD‐L1 (CPS and TPS) status among the three different areas were 71.6%, 74.6%, and 73.1%, respectively. The sensitivity and the specificity of the tumor surface for predicting the CD8 status of the whole tumor were high, especially in the NAC− group (both 1.0). The tumor surface also showed high capabilities for representing the whole PD‐L1 status, while yielding moderate positive predictive values (0.70). The ITH degrees and predictive capabilities did not differ according to NAC. Taken together, the ITH of CD8 and PD‐L1 differed among ESCC specimens, while not being markedly affected by NAC. The use of a biopsy specimen from the tumor surface might be feasible for TME evaluation.

Impacts of tumor microenvironment during neoadjuvant chemotherapy in patients with esophageal squamous cell carcinoma

“…16,18 In accordance with previous investigations, we herein showed the combination of T cells and PD-L1 status to be useful for stratifying survival outcomes. 10,12,27 No survival impact of CD8 + was detected in our cohort. A recent study underscored the significance of CD8 + status in ESCC patients, 18 although other studies have suggested modest survival impacts of CD8 + cells alone, especially in NAC-treated tumors.…”

“…46 A precise evaluation methodology for TME remains to be established in various specific malignancies, including ESCC. 19 Although we recently suggested that tumor surface information might reasonably be used for predicting the immune status of the whole tumor in ESCC, 27 TME evaluation of small tumors, such as early stage resections or biopsy samples, are easily affected by ITH. We evaluated only primary tumors.…”

“…Immunohistochemical studies were carried out as described previously. 27 Primary mAbs directed against CD4 + (clone SP35; Japan Tanner), CD8 + (clone C8/144B; Dako, Agilent Technology), CD20 + (L-26; Dako, Agilent Technology), Foxp3 + (clone 236A/E7; Abcam), HLA class I + (EMR8-5; Abcam), CD204 + (SRA-C6; Trans Genic Inc.), and PD-L1 + (PD-L1 22C3 pharmDx; Dako, Agilent Technology) were used (Table S1). Immunohistochemical staining was undertaken using a Dako Autostainer Link 48 platform (Dako, Agilent Technology) and the Ventana system on formalin-fixed, paraffin-embedded specimens.…”

“…Within the annotated area, the density of CD4 + , CD8 + , CD20 + , Foxp3 + , HLA class I + , CD204 + , and PD-L1 + -positive cells was estimated using HALO software (version 3.4; Indica Labs), as previously described. 27 The density of T cells was calculated by adding the CD4 + and CD8 + densities. The HALO immune cell algorithm (version 2.1) was used to detect the total density of CD8 + , CD4 + , CD20 + , and Foxp3 + -positive lymphocytes; the cytonuclear algorithm (version 2.0.9) to determine the density of cytoplasmic HLA class I + cells; the area quantification algorithm (version 2.1) to detect the proportion of CD204 + area.…”