The predictive value of preoperative neutrophil-lymphocyte ratio (NLR) on the recurrence of the local pigmented villonodular synovitis of the knee joint

Pigmented villonodular synovitis (PVNS) is a rare and benign disease of the synovial membrane characterized by abnormal synovium proliferation and hemosiderin deposition. According to previously reported statistics, the incidence rate is about 1.8/100000. It mostly affects the knee joint [1, 2], but can occurs in any synovial joint including the hip, ankle and elbow joint [3, 4]. The PVNS can be classified into the localized and diffuse forms (LPVNS and DPVNS) [5]. The etiology of the PVNS is still unknown yet while it was regarded as an inflammatory disorder in the past decades. Some risk factors have been recognized such as the trauma, chronic inflammation, and abnormal lipid metabolism [6]. Furthermore, PVNS would invade the adjacent bone and soft tissue leading to the bone errosion and the deformity of the involved joint [7, 8]. The high recurrence rate and metastasis risk of PVNS had also been reported [9], hence, up to now, PVNS has been considered to be a neoplastic-like disorder of the synovium, with synovitis as a secondary reaction in PVNS [10].

Many researchers has found that some hematological parameters like CRP [11], platelet volume [12], NLR [13] and PLR [14] are closely related with the outcomes of many diseases such as inflammatory disease, autoimmune disease and neoplastic disease [15]. Notably, NLR, PLR or LMR are new, simple and cost-effective predictors for prognosis. A meta-analysis conducted by Zhang J [16] found that the elevated NLR has a close relationship with the poorer overall survival of colorectal cancer (HR = 1.92 95%CI = 1.57–2.34; P < 0.00001). Kaida T et al. [17] also reported that PLR is an independent predictive factor of recurrence beyond the Milan criteria after liver resection for patients with hepatocellular carcinoma (odds ratio, 2.55; 95% confidence interval, 1.17–5.49; P = 0.018). However, there have been no reports regarding the relationship between NLR, PLR or LMR and the relapse of PVNS of knee joint, in addition, no quantitative parameters now can predict the recurrence of the PVNS of knee joint effectively.

In this study, we retrospectively reviewed the clinical characteristics, blood indexes, and recurrence of sixty cases diagnosed with LPVNS of knee joint in our department. The purpose of the present study is to explore the relationship between the NLR, PLR or LMR and the recurrence of the PVNS of knee joint. We hypothesized that some of these parameters could be used as new predictors for the recurrence of LPVNS of knee joint which were treated by arthroscopic surgery and adjuvant radiotherapy.

Sixty patients pathological diagnosed with PVNS were included (26 right knee, 34 left knee). The median age was 32 (range 14–75) years. On the MRI, the low signal intensity was presented on both T1 and T2 weighted images before the operation [18, 20]. The median duration of follow up was 52.8 (7–138) months and no complication of skin or wound infection was observed. Fourteen patients recurred and the median relapse time was 33.75 (20–51) months. Among the recurrent group, 3 patients underwent total knee arthroplasty and others received a second arthroscopy surgery.

Table 1 shows the clinical characteristics of the recurrent and non-recurrent group. The recurrent group showed higher NLR and CRP than non-recurrent group (P_NLR = 0.002, P_CRP = 0.04). There were no significant difference in other characteristics including PLR and LMR between the two groups (P_PLR = 0.23, P_LMR = 0.68).

The univariate and multivariate analysis were also performed with significant markers from ROC curves to determine which of them are independently associated with the diagnosis for relapse. The results indicates that NLR was significantly associated for prediction of relapse (odds ratio = 7.999, P = 0.017) (Table 2).

The optimal cutoff value that best distinguished recurrent from non-recurrent was determined at the maximum value, which was calculated by sensitivity + 1-specificity in the ROC curves. The ROC analysis of ESR, CRP, NLR, PLR and LMR showed the area under the curve (AUC) were 0.578 (95% CI = 0.443–0.704), 0.679 (95% CI = 0.545–0.793), 0.775 (95% CI = 0.649–0.873), 0.607 (95% CI = 0.473–0.731) and 0.537 (95% CI = 0.404–0.667) respectively (Fig. 2). Among the variables, the AUC of the NLR is the largest (AUC = 0.775, 95% CI = 0.649–0.873) and the optimal cut-off value is 2.42 for distinguishing the relapse which means when the NLR value is above 2.42 before treatment, the patient is more likely to have a relapse after the surgery. The Kappa test results (Kappa = 0.432, P = 0.001) and McNemar Chi-square test (P = 0.180) indicated that the NLR is valuable index for predicting the relapse.

The lyshoml and KSS score both improved in the two groups after the surgery though there is no significant difference before surgery and at the time of relapse in the recurrent group. Tables 3 and 4 shows that recurrent patients had poorer joint function than non-recurrent patients before the treatment did. In recurrent group, the knee functional score was higher at the time of relapse than that before operation while no significant difference was obtained. However, in non-recurrent group, most patients were satisfied with the greatly improved knee joint function and the difference in Lyshoml score and KSS score were statistically significant (P < 0.001) before surgery and at final follow up.

The PVNS is a locally aggressive, neoplastic-like disorder of synovial tissue that it would cause functional deterioration of the involved joint. The recurrence rate of the PVNS is as high as 10–50% for the residual diseased synovium after the simple synovectomy [21]. Isart A et al. [22] also reported a high recurrence of PVNS (61.5%, 8/13) after arthroscopic synovectomy. Moreover, it is almost impossible to resect the pathological synovium completely by open access or arthroscopic therapy [23]. So the high recurrence rate is still an intractable clinical issue despite many reports have claimed that adjuvant local radiotherapy after the synovectomy may be a salvage option [24]. What’s more, the non-specific symptoms of PVNS at its early stage often contribute to a delay in establishing a diagnosis, and the joint arthroplasty is the curative surgery in the terminal stage cases with severe bone erosion. So the early diagnosis of relapse is necessary in PVNS after the first treatment for its high recurrence rate. However, there has been no effective quantitatively marker for the relapse. The pathological examination is the gold standard and the MRI is the mostly used radiological tool for monitoring the relapse of PVNS. However, the MRI is qualitative and expensive and not available in some remote districts, additionally, the pathological examination always needs a second invasive procedure. Hence, finding an easy, non-invasive and cost-effective biomarker for relapse is a great challenge for surgeons.

The NLR can be obtained simply from neutrophil and lymphocyte counts. In addition, the NLR is cheaper relatively and it is obtainable in most medical institution for the blood routine examination is one of the routine preoperative test. Sever studies have demonstrated that the NLR、LMR and PLR are significantly associated with the outcome of many diseases. For example, Kucuk A et al. [25] reported that the NLR is obviously higher in Ankylosing spondylitis (AS) patients compared to controls. (NLR = 2.47 ± 1.33, 1.72 ± 0.47, respectively, P < 0.0001). Additionally, there is a significant difference between the severe AS disease activity and the mild AS disease activity (NLR = 2.72 ± 1.41, 2.20 ± 1.19, P = 0.001). The results of the ROC analysis (cutoff value = 1.91, sensitivity 69%, specificity 54%) also proved that NLR is a simple and inexpensive marker to indicate disease activity in patients with AS in daily clinical practice. Lin JP et al. [26] found preoperative LMR is an independent prognostic factor for GC which can improve the predictability of individual survival and recurrence of gastric cancer. Moreover, the predictive value of the NLR, LMR or PLR have been found in more and more diseases including the diabetes, hypertension, pancreatic cancer and hepatocellular carcinoma [27‐29].

To our knowledge, there have been no report about the NLR, LMR or PLR in predicting the recurrence of the knee PVNS. In our study, 60 patients diagnosed with LPVNS were enrolled. With respect to the gender ratio, our result (1.07, 31/29) was similar to those reported in Portugal (1.15) [30]. The BMI has no significant difference between two groups All participants received an arthroscopic synovectomy for the diagnosis of the LPVNS, and Xie et al. [15] found that no significant recurrence difference was identified between PVNS patients that were treated with open versus arthroscopic surgery (p = 0.78). In addition, the knee arthroscopic surgery, a minimally invasive surgery, is beneficial for fast recovery and shorter the hospitalization time. What’s more, the local adjuvant radiotherapy was applied to decrease the local relapse as much as possible. In a multicenter retrospective study, the recurrence rate in patients with knee PVNS is 24% (42/175) [2]. In our study, the recurrence rate is about 23.3% (14/60) which is consistent with the literatures. Among all hematologic indexes, Only CRP and NLR are significantly higher in recurrent patients (P_CRP = 0.04, P_NLR = 0.002). Other parameters like WBC, Neutrophils count and ESR are higher in recurrent group compared to controls, but don’t reach a statistical difference (P_WBC = 0.51, P_ESR = 0.208) which may implied that the PVNS is not a simple inflammatory disease. The results of ROC analysis and the multivariate analysis (Odds ratio = 7.99, 95% CI = 1.451–44.103, P = 0.017) indicated that the NLR may be a valuable marker for predicting relapse of knee PVNS after the treatment.

Our report suggests that the preoperative NLR is a valuable marker to predict the relapse of Knee PVNS treated by arthroscopic synovectomy combining radiotherapy. Our study indicated that these patients (preoperative NLR > 2.42) should be closely followed after the operation for the higher possibility of relapse. We hope this study could provide the orthopedic clinicians with a new method for predicting the postoperative recurrence of patients with PVNS.