The pretreatment Controlling Nutritional Status (CONUT) score is an independent prognostic factor in patients with resectable thoracic esophageal squamous cell carcinoma: results from a retrospective study

Despite significant improvements in the diagnosis and treatment of patients with esophageal cancer, their prognosis still remains poor due to its aggressive biological behavior [1]. Although surgical resection is the mainstay of treatment for local and locoregional disease in esophageal cancer, neoadjuvant treatment has been widely accepted as a means of improving the prognosis of esophageal cancer [2‐5]. Therefore, predicting prognosis using pretreatment clinical variables, but not operative and pathological variables, is important to improve the prognosis and to offer an optimal treatment strategy.

There is accumulating evidence that the presence of a systemic inflammatory response and malnutrition are associated with a worse prognosis in various malignancies [6‐9]. Recently, several inflammation-based markers, such as the platelet to lymphocyte ratio (PLR), the neutrophil to lymphocyte ratio (NLR), and the Glasgow Prognostic Score (GPS), have been reported to be prognostic factors in various malignancies, including esophageal cancer [10‐15]. The Controlling Nutritional Status (CONUT) score, which is calculated by the serum albumin concentration, the total peripheral lymphocyte count, and the total cholesterol concentration, was developed as a screening tool for early detection of poor nutritional status [16]. Use of the CONUT score has some advantages, such as simplicity and cost effectiveness, but there have been few reports on the relationship between the CONUT score and clinical outcomes in malignancies [17]. Therefore, the significance of the CONUT score in the treatment of esophageal cancer is still unknown.

The aim of this study was to elucidate the impact of the pretreatment CONUT score on survival compared with other inflammation-based markers (PLR, NLR, and GPS) in patients with resectable thoracic esophageal squamous cell carcinoma (ESCC).

The clinical data of consecutive patients who underwent subtotal esophagectomy with curative intent for resectable thoracic ESCC at Osaka City University Hospital (Osaka, Japan) between January 2000 and December 2014 were retrospectively reviewed. In this study, resectable thoracic ESCC was defined as patients without cT4 tumor and distant metastases on pretreatment examination. All patients were diagnosed with ESCC by biopsy before initial treatment. For reliable analysis, only thoracic ESCC patients who underwent two- or three-field lymphadenectomy and reconstruction using a gastric tube through the posterior mediastinum by cervical anastomosis were included. There was no uniform guideline for preoperative treatment until 2009; from 2009, neoadjuvant chemotherapy consisting of 5-fluorouracil/cisplatin or 5-fluorouracil/nedaplatin was administered for patients with clinical stage II/III in principle. Adjuvant chemotherapy was scheduled for patients with positive lymph node metastasis. Eight patients whose entire set of preoperative laboratory data was not available were excluded from this study. Ultimately, 185 patients were included. Forty-six patients received preoperative treatment; 39 patients received chemotherapy, 6 patients received chemoradiotherapy, and 1 patient received radiotherapy. This retrospective study was approved by the ethics committee at our institution and was conducted in accordance with the principles of the Declaration of Helsinki. Informed consent was obtained from all patients before treatment.

The pretreatment staging workup in principle included physical examination, laboratory tests, esophageal barium meal examination, upper GI endoscopy, enhanced computed tomography (CT) scans between the neck and upper abdomen, and positron emission tomography-computed tomography (PET-CT). On the basis of these examinations, tumor stage was assessed using the 6^th edition of the International Union Against Cancer [18]. Blood samples were obtained during the patients’ first visit to our institution before initial treatment. The CONUT score was calculated as described in Table 1. The PLR was calculated by dividing the platelet count by the lymphocyte count. The NLR was calculated by dividing the neutrophil count by the lymphocyte count. The GPS was constructed as follows: patients with both elevated C-reactive protein (CRP) (>1.0 mg/dl) and low albumin (<3.5 g/dl) were assigned GPS 2, those with elevated CRP (>1.0 mg/dl) or low albumin (<3.5 g/dl) were assigned GPS 1, and those with normal CRP (≤1.0 mg/dl) and normal albumin (≥3.5 g/dl) were assigned GPS 0.

Pretreatment variables, such as age, sex, body mass index (BMI), Eastern Cooperative Oncology Group performance status (PS), American Society of Anesthesiology score (ASA), tumor location, clinical TNM stage, and serum squamous cell carcinoma antigen (SCCA) levels were evaluated.

In the present study, the prognostic impacts of the CONUT score, PLR, NLR, and GPS were evaluated and compared in 185 patients with resectable thoracic ESCC. We found that only the CONUT score was an independent prognostic factor for OS and RFS, and it was superior to other inflammation-based markers in terms of predictive ability for prognosis before initial treatment.

The CONUT score was developed to assess nutritional status more easily and more objectively, having been validated in comparison with two other classical but slightly complicated assessment tools: the Subjective Global Assessment and the Full Nutritional Assessment [16]. The CONUT score is composed of the serum albumin concentration, total peripheral lymphocyte count, and total cholesterol concentration. The serum albumin concentration is known as a reliable indicator of nutritional status and systemic inflammation [20]. Total peripheral lymphocytes, which play an important role in the immune response to the tumor, are known to indicate the immunological and nutritional status. Total cholesterol concentration is known as an indicator of a patient’s caloric reserves [21]. Thus, a higher CONUT score could reflect not only malnutrition, but also systemic inflammation and an impaired immune response. Indeed, the high-CONUT group was significantly associated with lower BMI, high-PLR, high-NLR, and GPS1/2.

Recently, Hirahara et al. [22] first reported that the CONUT score was an independent predictor of cancer-specific survival in patients who underwent curative thoracoscopic esophagectomy for esophageal cancer. They also demonstrated that a high CONUT score was significantly associated with high NLR, and NLR was shown not to be a significant prognostic factor, consistent with the present results. The present study advances this previous study by determining the cut-off value using statistical methods, showing that the CONUT score is an independent prognostic factor for both OS and RFS and by comparing it with several inflammation-based markers.

The difference between the CONUT score and other inflammation-based markers is that the CONUT score includes the total cholesterol concentration in its calculation. Cholesterol is an essential component of the cell membrane that is involved in numerous biochemical pathways potentially correlated with cancer initiation and progression and the immune response. Several epidemiological studies and case control studies have demonstrated that a lower serum total cholesterol level was associated with increased mortality from several cancers [23‐29]. Similarly, in the present study, a lower serum total cholesterol level was significantly associated with poorer survival. However, it remains uncertain whether hypocholesterolemia is a result of tumor progression or induces tumor progression. As a potential hypothesis to explain the relationship between hypocholesterolemia and cancer, a lowered serum cholesterol is attributed to the increased consumption of cholesterol needed for growth in cancer cells, which may indicate higher activity and malignancy of the tumor [25, 30]. This hypothesis could be supported by the fact that serum cholesterol levels increase after curative surgery and then decrease with cancer recurrence [31]. On the other hand, Calleros et al. [32] demonstrated that chronic cholesterol depletion induces NFkB activation, which could promote proliferation of malignant tumor cells. Muldoon et al. [33] reported that hypocholesterolemia was significantly more associated with fewer circulating lymphocytes, total T cells, and CD8+ cells than hypercholesterolemia. It was also reported that cholesterol increases the antigen-presenting function of monocytes [34]. Thus, a low serum total cholesterol level may contribute to a poorer prognosis by affecting intracellular signaling and impairing the immune system against tumor spread. Inclusion of the serum total cholesterol level in its calculation may be one of the reasons why the CONUT score is able to predict patients with a poorer prognosis more sensitively than other inflammation-based markers.

As in other malignancies, several studies have shown that PLR, NLR, and GPS were independent prognostic factors in esophageal cancer [11, 35]. In contrast, several studies did not show PLR and NLR to be independent prognostic factors [36‐38]. In the present study, although PLR, NLR, and GPS were found to be predictive factors for OS and RFS on univariate analyses, multivariate analysis did not show these markers to be independent prognostic factors. These discrepancies may be due to the differences in sample size, follow-up periods, and cut-off values, which vary by method of determination and the population. Because the method to determine the optimal cut-off value has still not been fully established, various cut-off values have been used in previous studies [11, 14, 39]. In this study, the cut-off values to predict survival were determined from the maximal Youden indices based on the results of time-dependent ROC curve analyses [19, 40]. The present methodological approach seems to have advantages in terms of objectivity, but its validity has not been evaluated. Further studies are needed to establish the best method to determine cut-off values. The present study focused on resectable esophageal cancer patients who may have less systemic inflammation and better nutritional status than those with T4 and metastatic disease. Therefore, the proportions of high PLR and high NLR patients were lower than in previous reports [11, 41]. It might be important to identify an optimal cut-off value based on similar populations, such as tumor stage.

It is noteworthy that the CONUT score calculated before initial treatment was an independent prognostic factor, contributing to making an individual treatment strategy. Although the effects of perioperative nutritional intervention on the long-term outcome in patients with malnutrition due to malignant disease have not been confirmed, numerous studies have reported that perioperative nutritional intervention improved tolerance for anticancer treatment and reduced postoperative complications, which may contribute to improving the prognosis [42‐44]. However, the definition and assessment of malnutrition in previous studies have not been unified. The CONUT score may be used as an index for selection of patients who need nutritional intervention and for evaluation of nutritional management in esophageal cancer treatment.

In the present study, the high-CONUT group was significantly associated with primary cancer death and poorer RFS. Similarly, Iseki et al. [17] and Hirahara et al. [22] reported that a higher CONUT score was an independent predictive factor for cancer-specific survival in colorectal cancer and esophageal cancer, respectively. These findings suggest that poorer nutritional status identified by the CONUT score may be involved in poorer tolerability for anticancer treatment and the growth of micrometastatic and residual cancer cells, which results in a worse prognosis. Patients with a high CONUT score may be candidates for not only nutritional intervention, but also more intensive multimodal treatment in resectable ESCC.

The present study has some limitations. First, this was a retrospective study conducted at a single institution, and the number of cases was limited. Second, potential factors that affect inflammation-based and nutritional markers, such as comorbidities and medications, could not be excluded. Third, there was heterogeneity in the neoadjuvant and adjuvant treatments in this study. However, in the subgroup analysis with or without neoadjuvant treatment, the high-CONUT group had significantly worse OS and RFS in both subgroups. Within these limitations, the present study demonstrated that the CONUT score is a promising prognostic factor with a better predictive value than PLR, NLR, and GPS in patients with resectable thoracic ESCC. Large-scale prospective validation studies are needed to confirm these findings.

The CONUT score was found to be an independent predictor of OS and RFS before treatment, and it was superior to PLR, NLR, and GPS in terms of predicting prognosis in patients with resectable thoracic ESCC. The estimation of the CONUT score is inexpensive and easily available from laboratory data in daily clinical practice. We suggest that the CONUT score should be calculated routinely before initial treatment, and it could be a useful indicator for pretreatment nutritional management in esophageal cancer.