1. Introduction

Esophageal cancer is a malignant tumor with high incidence and mortality rates, with over 600,000 new cases globally each year.[1] In China, it ranks sixth and fifth in terms of incidence and mortality, respectively, among all malignant tumors.[2] Furthermore, the pathological types of esophageal cancer differ significantly among different regions, with adenocarcinoma being more prevalent in European and American countries, while squamous cell carcinoma in China.[3]

Surgical treatment is the primary therapeutic approach for early-stage esophageal cancer, which can significantly improve the short-term survival rate of patients.[4] For those who are unwilling or unable to undergo surgery, radical chemoradiotherapy (CRT) is the standard alternative.[5–7] Increasing evidence suggests that systemic inflammatory responses play a vital role in the occurrence and development of tumors and are closely associated with tumor proliferation, invasion, and metastasis.[8] Changes in tumor-related inflammatory cells reflect the degree of the body’s inflammatory response to the tumor: the higher the levels of inflammation, the poorer the prognosis.[9,10]

In recent years, clinical trials have reported that systemic inflammatory markers, such as the systemic immune-inflammation index,[11] advanced lung cancer inflammation index (ATI),[12] neutrophil-to-lymphocyte ratio (NLR)[13] and platelet-to-lymphocyte ratio (PLR),[14] are important prognostic indicators for patients with malignant tumors such as colorectal cancer, lung cancer, and cervical cancer.

In particular, PLR can be referred to in predicting the prognosis of patients with esophageal squamous cell carcinoma (ESCC), most studies on the predictive role of PLR tend to focus on the impact of PLR levels before treatment on prognosis.[14–16] However, there are few studies on the impact of changes in PLR levels in different treatment stages on the prognosis of cancer patients. At the same time, its use is controversial because its accuracy and optimal cutoff value remain unclear.

Therefore, this study retrospectively analyzed the clinicopathological data of 168 ESCC patients undergoing CRT, in order to explore the influence of dynamic changes in PLR at different treatment stages (before and after treatment) on the prognosis of ESCC patients undergoing radiotherapy (RT) and chemotherapy. The accuracy of PLR in predicting prognosis was also evaluated through subgroup analysis.

2. Materials and methods 2.1. Study object

This retrospective study was approved by the Ethics Committee and Institutional Review Board of Guangyuan Central Hospital (No. GYZXLL202310). A total of 168 patients with ESCC who received CRT from January 2012 to December 2018 were included. The 7th edition of the American Joint Committee on Cancer TNM staging was adopted. The inclusion criteria were as follows: age > 18 years; KPS score ≥ 70 points; pathologically confirmed squamous cell carcinoma of the esophagus; patients receiving curative RT; patients with laboratory and clinical data; and no distant tumor metastasis. The exclusion criteria included previous surgery history, incomplete follow-up information, and any other primary tumors aside from esophageal cancer.

2.2. Treatment

Three-dimensional conformal radiotherapy or intensity-modulated RT was performed. Bilateral lung V20 ≤ 20%, bilateral lung V5 < 50%, heart V30 ≤ 30%, and the maximum spinal cord dose ≤ 45.0 Gy. Simultaneous chemotherapy was given at the same time, using platinum-based combined chemotherapy for 1 to 5 cycles: (1) paclitaxel + cisplatin, (2) paclitaxel + carboplatin, and (3) docetaxel + cisplatin. For patients who cannot tolerate intravenous chemotherapy, 1–3 cycles of oral chemotherapy with S-1/capecitabine are given.

2.3. Observation indicators and follow-up

A follow-up evaluation was conducted every 3 months in the first year, then every 6 months in the next 2 years, and then at the end of each year or until death. All patients were followed up through outpatient examinations and telephone calls until December 2022.

2.4. PLR calculation method

PLR = platelet/lymphocyte, pretreatment PLR is the blood test index 1 week before treatment; mid-treatment PLR (mid-PLR) is the lowest level of blood test index during the treatment process; posttreatment PLR (Post-PLR) is the blood test index 1 week after treatment.

2.5. Statistical analysis

R 4.2.1 software was used for statistical analysis. The continuous variables conforming to the normal distribution were compared by t test or analysis of variance, and the χ2 test was used to compare count data. Log-rank test and optimal stratification were performed to determine the optimal cutoff value of continuous covariate PLR.[17,18] The best cutoff value of PLR was used as reference in dividing the patients into the low PLR and high PLR groups (Figs. 1 and 2). For survival analysis, the Kaplan–Meier method and Log-rank test were performed. For univariate and multivariate analyses, Cox risk model was used. Test level α was 0.05, and P < .05 was set for statistical significance.

Figure 1.:

Pre-PLR using the Log-rank test to calculate optimal stratification cutoffs for continuous covariates.

Figure 2.:

Post-PLR using the Log-rank test to calculate optimal stratification cutoffs for continuous covariates.

3. Results 3.1. Patient and tumor characteristics

The median age of the 168 patients was 63 years (34–86 years), including 131 males (78.0%) and 37 females (22.0%). Of the 168 patients, patients (57.7%) had smoking history, while 71 (42.3%) did not have smoking history; 91 (54.2%) had drinking history, while 77 patients (45.8%) did not have drinking history. A total of 68 patients (40.5%) had a tumor length < 5 cm, while 100 (59.5%) had a tumor length ≥ 5 cm. 6 patients (4.2%) were stage II, while 137 (95.8%) were stage III. In terms of RT dose, 26 patients (15.4%) received < 60 Gy, 142 (84.6%) received ≥ 60 Gy, 142 patients (84.6%) received chemotherapy, while 26 (15.4%) did not receive chemotherapy (Table 1). With regard to the changes in PLR levels before and after treatment, results of the paired analysis showed that the levels of Post-PLR in patients significantly decreased to varying degrees compared with those before treatment (P < .05, Fig. 3). The distribution of background variables for survival stratification is shown in Table 2. Compared with the patients who died, the surviving patients smoked and drank less, had shorter tumor length, were in earlier T stage, and had lower PLR, with significant differences in T stage, pre-PLR, and post-PLR Lower (P < .05, Table 1).

Table 1 - Clinicopathological features of 168 patients with ESCC. Characteristics N Alive (n = 70) Death (n = 98) P value Gender, n (%) .176  Female 37 (22.0%) 19 (11.3%) 18 (10.7%)  Male 131 (78.0) 51 (30.4%) 80 (47.6%) Age (years), n (%) .249  <65 92 (54.8%) 42 (25%) 50 (29.8%)  ≥65 76 (45.2%) 28 (16.7%) 48 (28.6%) KPS score, n (%) .169  70 5 (3%) 0 (0%) 5 (3%)  80 68 (40.5%) 27 (16.1%) 41 (24.4%)  90 94 (56.0%) 43 (25.6%) 51 (30.4%)  100 1 (0.6%) 0 (0%) 1 (0.6%) Smoking history, n (%) .444  No 71 (42.3%) 32 (19%) 39 (23.2%)  Yes 97 (57.7%) 38 (22.6%) 59 (35.1%) Alcohol history, n (%) .063  No 77 (45.8%) 38 (22.6%) 39 (23.2%)  Yes 91 (54.2%) 32 (19%) 59 (35.1%) Location, n (%) .459  Cervical 6 (3.6) 3 (1.8%) 3 (1.8%)  Upper thoracic 34 (20.2%) 16 (9.5%) 18 (10.7%)  Middle thoracic 56 (33.3%) 22 (13.1%) 34 (20.2%)  Lower thoracic 64 (38.1%) 28 (16.7%) 36 (21.4%)  Abdomina 8 (4.8%) 1 (0.6%) 7 (4.2%) Tumour length (cm), n (%) .034  <5 68 (40.5%) 35 (20.8%) 33 (19.6%)  ≥5 100 (59.5%) 35 (20.8%) 65 (38.7%) T-stage, n (%) <.001  T2 22 (13.1%) 19 (11.3%) 3 (1.8%)  T3 77 (45.8%) 34 (20.2%) 43 (25.6%)  T4 69 (41.1%) 17 (10.1%) 52 (31%) N-stage, n (%) .130  N0 6 (3.6%) 5 (3%) 1 (0.6%)  N1 54 (32.1%) 23 (13.7%) 31 (18.5%)  N2 79 (47.0%) 33 (19.6%) 46 (27.4%)  N3 29 (17.3%) 9 (5.4%) 20 (11.9%) RT dose (Gy), n (%) .349  <60 26 (15.4%) 13 (7.7%) 13 (7.7%)  ≥60 142 (84.6%) 57 (33.9%) 85 (50.6%) Chemotherapy, n (%) .718  No 26 (15.4%) 10 (6%) 16 (9.5%)  Yes 142 (84.6%) 60 (35.7%) 82 (48.8%) Pre-PLR, n (%) .022  <195.7 132 (88.5%) 61 (36.3%) 71 (42.3%)  ≥195.7 36 (21.5%) 9 (5.4%) 27 (16.1%) Post-PLR, n (%) .032  <403.6 99 (59.0%) 48 (28.6%) 51 (30.4%)  ≥403.6 69 (41.0%) 22 (13.0%) 47 (28%)

ESCC = esophageal squamous cell carcinoma, PLR = platelet-to-lymphocyte ratio, RT = radiotherapy.

Table 2 - Univariate and multivariate analysis of prognostic factors. Characteristics Total (N) Univariate analysis Multivariate analysis Hazard ratio (95% CI) P value Hazard ratio (95% CI) P value Gender, n (%) .219  Male 37 Reference  Female 131 1.365 (0.818–2.277) .234 Age (years), n (%) .068  <65 92 Reference Reference  ≥65 76 1.450 (0.975–2.158) .067 1.736 (1.129–2.669) .012 KPS score, n (%) .368  70 5 Reference  80 68 0.437 (0.172–1.108) .081  90 94 0.404 (0.161–1.014) .054  100 1 0.714 (0.083–6.130) .758 Smoking history, n (%) .403  No 71 Reference  Yes 97 1.188 (0.792–1.781) .405 Alcohol history, n (%) .089  No 77 Reference Reference  Yes 91 1.416 (0.944–2.122) .092 1.622 (1.050–2.508) .029 Location, n (%) .809  Cervical 6 Reference  Upper thoracic 34 1.211 (0.356–4.113) .759  Middle thoracic 56 1.264 (0.388–4.122) .698  Lower thoracic 64 1.245 (0.383–4.053) .715  Abdomina 8 2.045 (0.527–7.931) .301 Tumour length (cm), n (%) .010  <5 68 Reference Reference  ≥5 100 1.722 (1.130–2.623) .011 1.239 (0.797–1.925) .340 T-stage, n (%) <.001  T2 22 Reference Reference  T3 77 5.719 (1.773–18.453) .004 5.039 (1.541–16.477) .007  T4 69 14.447 (4.458–46.814) <.001 12.483 (3.719–41.896) <.001 N-stage, n (%) .007  N0 6 Reference Reference  N1 54 4.359 (0.594–31.954) .148 1.867 (0.247–14.115) .545  N2 79 5.118 (0.705–37.142) .106 1.720 (0.228–12.960) .599  N3 29 9.599 (1.284–71.770) .028 1.480 (0.182–12.028) .714 RT dose (Gy), n (%) .638  <60 26 Reference  ≥60 142 1.148 (0.640–2.058) .644 Chemotherapy, n (%) .536  No 26 Reference  Yes 142 0.841 (0.492–1.439) .528 Pre-PLR .001  <196.15 132 Reference Reference  ≥196.15 36 2.165 (1.383–3.387) <.001 1.716 (1.069–2.756) .025 Post-PLR .003  <403.6 99 Reference Reference  ≥403.6 69 1.827 (1.228–2.718) .003 1.664 (1.106–2.503) .015

CI = confidence interval, ESCC = esophageal squamous cell carcinoma, HR = hazard ratio, PLR = platelet-to-lymphocyte ratio, RT = radiotherapy.

Figure 3.:

Changes of PLR in different treatment stages. (A) The levels of pre-PLR and post-PLR. (B) Paired analysis of pre-PLR and post-PLR (PLR before treatment: pre-PLR; PLR after treatment: post-PLR).

3.2. Univariate and multivariate analyses

Results of the univariate analysis showed that tumor length, T stage, N stage, pre-PLR, and post-PLR could predict the prognosis of patients with ESCC (all P < .05, Table 2). On multivariate analysis, age (hazard ratio [HR] = 1.736; 95% confidence interval (CI) = 1.129–2.669; P = .012), drinking (HR = 1.622; 95%CI = 1.050–2.508; P = .029), T stage (HR = 12.483; 95%CI = 3.719–41.896; P < .001), pre-PLR (HR = 1.716; 95%CI = 1.069–2.756; P = .025), post-PLR (HR = 1.664; 95%CI = 1.106–2.503; P = .015) were independent factors of the prognosis of patients with ESCC (Table 2).

3.3. Demographic and clinical characteristics of cancer patients stratified by PLR at different treatment stages

Using the Log-rank test, the optimal stratification was used to resolve the optimal cutoff value of the continuous covariate, and the data on both sides of the PLR value were distinguished to obtain the best difference. Using this method, the pre-PLR cutoff value was 195.7, and the patients were divided into low pre-PLR group (PLR < 195.7; n = 132) and high pre-PLR group (PLR ≥ 195.7; n = 36). Pre-PLR was associated with sex, T stage, and N stage (all P < .05, Fig. 1, Table 3). Post-PLR cutoff value was 403.6: low post-PLR group (PLR < 403.6; n = 132) and high post-PLR group (PLR ≥ 403.6; n = 36; Fig. 2, Table 4).

Table 3 - Demographic and clinical characteristics of cancer patients stratified by pre-PLR. Characteristics Total (N) Pre-PLR < 195.7 (n = 132) Pre-PLR ≥ 195.7 (n = 36) P value Gender, n (%) .674  Male 131 102 (60.7%) 29 (17.3%)  Female 37 30 (17.9%) 7 (4.2%) Age (years), n (%) .627