The Relationship Between Evaluation Methods for Chemotherapy-Induced Peripheral Neuropathy

Numbness and pain are currently evaluated using subjective methods such as the visual analogue scale (VAS). PainVision (PV) is an analytical instrument that was designed to quantitatively assess sense perception and nociception in patients. Chemotherapy-induced peripheral neuropathy (CIPN) is one of the most important adverse events that renders prolonged chemotherapy difficult. To assess the features of CIPN, we aimed to compare PV methods with existing methods. A total of 73 patients received oxaliplatin for metastatic colorectal cancer. Registered patients included 37 men and 36 women in the range of 37 to 89 years (median 70). CIPN was evaluated a total of 483 times (median per patient six times). Our study examined the correlation between evaluation methods of CIPN using VAS and PV, respectively. The average VAS (hand), VAS (foot) and PV scores of CIPN were 18.4 (range: 0–100), 23.8 (range: 0–100), and 24.7 (range: 0–496), respectively. VAS (hand), VAS (foot), and FACT/GOG-NTX (NTX2, NTX4 and NTX8) were significantly correlated with PV. PV showed no correlation with a Disk-Criminator or the monofilament test used as a quantitative evaluation. The evaluation of CIPN is complex, and further improvement is required for evaluation with PV.

A strong positive correlation was found between VAS (hand) and VAS (foot) scores (r = 0.798) (Fig. 1). Each data point represents one assessment from a single patient. The average value obtained by subtracting the VAS (hand) from VAS (foot) was 5.38 (SD: 18.32), which was not significant in the t-test (p < 0.001). This finding indicates that the average VAS (foot) value was higher than the VAS (hand) value.
A scatter plot of PV and FACT/GOG-NTX, a scatter plot of PV and VAS, a scatter plot of PV, and Disk-Criminator or monofilament are shown in Figs. 2, 3 and 4, respectively. In Fig. 2, the horizontal axis represents a score of 0-4 for each FACT/GOG-NTX question item, and the vertical axis represents the PV score. Despite evaluating the same symptoms, neither show a strong correlation. These figures do not take into account intraindividual variability. Thus, in Table 3, the results of hierarchical mixed models, including random individual effects and fixed effects of age and sex are shown. VAS (hand), VAS (foot), NTX 2, NTX 4 and NTX 8 were significantly associated with PV. There were no significant associations of the Disk-Criminator ™ and monofilament methods with PV scores ( Table 3). The associations between the repeated measurements of changes www.nature.com/scientificreports www.nature.com/scientificreports/ from the initial evaluations of VAS (hand), VAS (foot), Disk-Criminator, monofilament, and FACT/GOG-NTX, and those in PV were analyzed using a hierarchical mixed model to determine the best method for detecting increased CIPN symptoms when patients received additional chemotherapy ( Table 4). The lowest P-value was VAS (foot), which correlated best with the amount of change over time. In Fig. 5, the lower the Disk-Criminator score, the higher the PV score, VAS (hand), and NTX1 values. Those with a Disk-Criminator score of 9 did not have high PV scores, VAS (hands), and NTX1 values. Similar findings were also observed for the monofilament score (Fig. 6).

Discussion
VAS is one of the most common methods used to assess pain 24 . VAS is a method of grading pain currently experienced by patients compared with the worst imaginable pain 25 . Because of its ease of use, VAS has become a popular tool to quantify pain relief and pain intensity. VAS is an effective and reliable means to assess pain, depression, anxiety, and mood 24 . VAS tends to focus only on pain intensity, with an increased risk of over-simplification of the experience 26    www.nature.com/scientificreports www.nature.com/scientificreports/ objectively evaluating the extent of CIPN, enabling the identification of asymptomatic peripheral neuropathy before onset 28 . In our study, VAS (hand), VAS (foot), NTX 2, NTX 4 and NTX 8 significantly correlated with PV in the analyses using a regression model with random effects of individual IDs adjusted by sex and age. Sex differences in the experience of clinically and experimentally induced pain are widely reported 29 , pain sensitivity is also thought to decrease with increasing age. However, as shown in Fig. 3, if intraindividual variability is not considered, the correlation coefficient between PV and VAS decreases. The Disk-Criminator and monofilament results did not associate with PV results. This discrepancy is because PV and VAS were used to assess pain, whereas the Disk-Criminator was used to measure spatial acuity in tactile sensations and monofilament was used to measure tactile sensitivity. Prior research has shown that numbness and tingling symptoms are correlated, but they are not necessarily correlated with pain 30 .
Several studies have analyzed the sensitivity to change over time in each of the proposed CIPN measures, in an attempt to test sensitivity to change over time [31][32][33][34] . Table 4 shows the association between repeated measures of changes from the initial evaluations of VAS (hand), VAS (foot), Disk-Criminator, monofilament, and FACT/ GOG-NTX, and those in PV. This result suggested that VAS is the most representative measure of changes in neuropathy over time. In Table 3, there was a significant difference in NTX8, but as shown in Table 4, the significant difference disappeared. NTX8 indicates a response of "I have trouble buttoning buttons, " and we posit that the patient was able to learn the movement with time.  www.nature.com/scientificreports www.nature.com/scientificreports/ PV is used for the quantification of the intensity of pain. In clinical practice, this method is used not only for chronic pain such as fibromyalgia 35 and lower back pain due to spondylolisthesis 36 but also for acute pain caused by the removal of wound dressings 37 . Previous studies have shown that PV is a useful device that can quantitatively evaluate pain in various fields [38][39][40] . Patient-reported outcomes of CIPN related symptoms should always be included in clinical trials 41 . Measures of clinician-rated neuropathy signs and function measures are also encouraged. Although an evaluation method that can quantify CIPN enables interindividual and intraindividual comparison, Sato et al. reported that there was no significant difference between PV and CTCAE grades in the evaluation of CIPN 42 . If the severity of CIPN and PV do not correlate, PV cannot be used to evaluate clinical trials aimed at improving CIPN. Although PV significantly correlated with VAS (hand), VAS (foot), NTX2 and NTX4 (Table 3), it does not appear in Fig. 3 to be correlated. Therefore, it is necessary to make improvements that are also correlated visually. Because PV was developed for pain assessment, the following four conditions are considered as speculations to improve the evaluation of CIPN: (1) Measurement by simultaneous stimulation of multiple parts; (2) Machine body and software corresponding to increase and decrease of stimulus in one measurement; (3) the optimal stimulation wave; (4) the ease of input for stimulus detection. Clinical trials for the prevention and treatment of CIPN require the identification of optimal outcome measures to define the CIPN phenotype and the setting of parameters that lead to the evaluation of clinically relevant effects 43 . If these four conditions are satisfied, the correlation coefficient seems to rise further.
This study has a limitation. We should have reported the change in each CIPN measure vs. time and the associated factors. Recently, oxaliplatin has been stopped before CIPN has developed with the spread of the Stop & Go strategy 44 , and oxaliplatin has been reduced or suspended immediately after CIPN has developed. In addition, there are individual differences in the timing of CIPN. By these two points, we could not report change in each CIPN measure vs. time and the associated factors. The associations between the repeated measures of changes from the initial evaluations were investigated as an alternative (Table 4). To the best of our knowledge, there has been no previous study regarding the correlation between PV and other assessment in CIPN patients. We believe that the effect of the drug for CIPN should be evaluated quantitatively. Further research and effort are needed to improve the evaluation of CIPN by PV.

conclusions
Evaluation of CIPN is complex because numerous factors are involved. To apply quantitative evaluation methods to CIPN clinical trials, PV requires various improvements.

Methods
Study design. This study was approved by the Institutional Review Board of Fukuoka University Hospital (No. 13-4-7) and was performed between April 2014 and December 2015. Seventy-three patients with histologically confirmed metastatic colorectal adenocarcinoma, and treated with oxaliplatin as the first line of chemotherapy, were enrolled in the study. Patients exhibiting mental health issues that rendered the concepts of PV impossible to understand were excluded from this study. Patients who had peripheral neuropathy or musculoskeletal pain that could interfere with the measurement of quantitative pain before chemotherapy were also excluded. Informed consent was obtained from all patients before participation in this study. All methods were implemented according to the Declaration of Helsinki. This study included a different patient cohort than our previously reported studies 19,20,23 . www.nature.com/scientificreports www.nature.com/scientificreports/ CIPN was defined using the National Cancer Institute Common Terminology for Adverse Events 45 . The measurement started from the second cycle and was performed before administration. Measurements are recorded continuously every 3 weeks in line with chemotherapy but are halted when chemotherapy is postponed because of adverse events. During treatment with oxaliplatin, measurements were recorded until the patient refused further measurement, and all measurements were analysed. VAS and the functional assessment of cancer therapy/Gynecologic Oncology Group -Neurotoxicity (Fact/GOG-NTX). VAS is a commonly used method for assessing the fluctuation of pain  www.nature.com/scientificreports www.nature.com/scientificreports/ intensity. Patients are instructed to indicate the perceived pain intensity by marking on a 100-mm horizontal line labelled "0 (no pain)" at the left end and "100 (worst imaginable pain)" on the right end (Fig. 7A). VAS was used to assess chronic CIPN subjectively before each cycle of chemotherapy. The patient was instructed to consider only neuropathic pain present on the day of the measurement.
The FACT/GOG-NTX is an 11-item subscale for evaluating symptoms associated with chemotherapy-induced neuropathy (Fig. 7B). We examined the correlation between these 11 items and PV.  www.nature.com/scientificreports www.nature.com/scientificreports/ PainVision PS-2100. PV was developed as a medical device that evaluates the strength of pain using a numerical value 46 . The measurement principle of the system is to compare a unique electrical stimulation with the pain experienced by the patient 19,37 . A painless electrical stimulation whose intensity is equivalent to the intensity of the pain experienced by the patient is applied, and the current value of this electrical stimulation is defined as "pain-compatible electrical current. " The patient's threshold for the electrical stimulation is defined as the "current perception threshold" which is intended to eliminate inter-individual variability. With these two values, pain intensity is defined by the following equation: -Pain intensity (pain compatible electrical current current perception threshold)/ current perception threshold 100 An electrode is mounted on the inside surface of the forearm. An electrical current is made to flow (50 Hz; 0-150 µA RMS; pulse width: 0.3 ms), and the stimulation is strengthened 19,22 . The patient is instructed to press a button the first time she/he perceives this stimulus; the current at this point is defined as the "minimum perceived current" value. As the stimulation current is increased, the patient is instructed to press the switch when they feel that the intensity of the stimulation current is equivalent to the intensity of the pain they are experiencing. The current is defined at this point as the "pain-equivalent current" value. Using the obtained values, "pain intensity" is calculated using the above formula. In the absence of pain, the value is 0 and increases according to the degree of pain. There is no upper limit. Each measurement is easily completed in a few minutes. PV was used to assess symptoms related to chronic CIPN subjectively before each cycle of chemotherapy.
the Disk-criminator test. The Disk-Criminator is a two-point discrimination (TPD) measuring device 47 (Fig. 7C). The TPD method is a method that is completed in a shorter time than the nerve conduction test, is less painful, practical, cost-effective, and more easily applicable 48 . The method was performed as previously reported 49 . The Disk-Criminator has nine levels of discrimination, the first being 0 mm, or 1 point, whereas the  www.nature.com/scientificreports www.nature.com/scientificreports/ rest were 2 points, with distances between the 2 points of 1 mm to 9 mm. For 2-point discrimination testing, patients were asked to respond with the number (1 or 2) they felt most accurately indicated the stimulus. The patient was blinded, and the hand immobilized. The tester applied just enough pressure to depress the ventral side of index finger directly below the instrument, and the points contacted the skin simultaneously. The placement of 1 or 2 points was randomly mixed. Each subject was assessed three times on each of the nine distances on the Disk-Criminator. The number of correct responses was the 2-point discrimination score. Measurements were taken directly before each treatment cycle.
The monofilament test. The monofilament test is an easy-to-use, inexpensive, and portable test for evaluating the loss of protective sensation and is recommended by several practical guidelines to detect peripheral neuropathy 50,51 . The test was performed using a Semmes-Weinstein aesthesiometer (Research Design, Inc., Houston, TX, USA) (Fig. 7D). The filament contacts the ventral side of the patient's index finger. With a loss of sensation, the patient cannot detect the presence of the filament. The higher the value of the monofilament, the stiffer and harder it is to bend. Four monofilaments used to diagnose peripheral neuropathy are the 2.83/0.07 g (Green), 3.61/0.4 g (Blue), 4.31/2.0 g (Purple) and 4.56/4.0 g (Red). The monofilaments were applied slowly and precisely to the skin of the finger in the same fashion for the same amount of time for each test. It was pressed to produce a slight bend. Every trial involved touching the patient and then recording whether the patient reported, "Yes, I was touched" or "No, I was not touched. " Each subject was assessed three times with each of the monofilaments. The smallest monofilament color that the patient could feel was entered. Measurements were taken directly before each treatment cycle.
Statistical analyses. Data were analysed using SAS Version 9.4 (SAS Institute, Cary, North Carolina, USA).
To investigate the reliability of the device, the quantified pain degree score was evaluated twice. Each measurement was performed twice, and the average value was used. Data are presented as the mean ± standard deviation (SD), median (interquartile range 25-75%), or the number of participants (percentages). The associations of VAS (hand), VAS (foot), Disk-Criminator, monofilament, and FACT/GOG-NTX with PV were assessed using hierarchical mixed models including random individual effects with or without fixed effects of age and sex. The associations between the repeated measures of changes from the initial evaluations in VAS (hand), VAS (foot), Disk-Criminator, monofilament, and FACT/GOG-NTX, and those in PV were evaluated using a hierarchical mixed model. P values less than 0.05 were considered statistically significant.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.