Grading of chemotherapy-induced peripheral neurotoxicity using the Total Neuropathy Scale

Patients and methods.

Following Institutional Review Board approval and patients’ written informed consent, 60 women with squamous cervical carcinoma were examined during treatment with TP (paclitaxel 175 mg/m² + cisplatin 75 mg/m²) or TIP (TP + iphosphamide 5 mg/m²) chemotherapy.

According to TNS (table 1), sensory, motor, or autonomic symptoms were assessed by interviewing the patients. The neurologic examination was based on the standard evaluation of strength, deep tendon reflexes (DTR), and pin and vibration sensibility. For the latter, the patient was asked to report the subjective perception of cessation of vibration while the 128 Hz tuning fork was gradually changing from the maximal vibration (score 0 if perceived) to nearly null (score 8). Nerve conduction studies were performed with a Medelec Premiere Plus electromyograph (Vickers Medicals, Woking, UK) using standard methods.10 Vibration detection threshold (VDT) was assessed by a trained examiner using a Vibrameter type IV device (Somedic AB, Stockholm, Sweden) at the first metatarsal bone using the method of limits (i.e., stimuli increased or decreased in a continuous manner until sensation occurred or disappeared and, finally, until the same value was obtained with the two different modalities). The neurophysiologic and VDT normal reference values were previously determined in the Department of Neurology in age-matched subjects (see table 1).

CIPN was also assessed by a single examiner (G.C.) using the NCI-CTC 2.0, ECOG, and Ajani scores (see table 1). Finally, a reduced version of TNS (TNSr) was calculated (see table 1), and correlations with the common toxicity scale results were recalculated (Spearman test for nonparametric data, significant level p < 0.05); the 95% CL were also calculated using PRISM software (GraphPad Software Inc., San Diego, CA).

Results.

A total of 110 examinations were available for the correlation study. No patient was graded more than 0 in the motor NCI-CTC 2.0 scale and, accordingly, no motor symptoms were reported by any of the patients. However, in some patients, mild distal weakness was observed at the neurologic examination. No autonomic symptoms were referred by the patients.

The results of the correlation study between TNS and sensory common toxicity scores are reported in table 2.

The overall TNS score correlated very closely with NCI-CTC 2.0, Ajani, and ECOG scores. The correlation between TNS and the selected neurotoxicity scales was calculated also after stratification of the entire population into two subgroups according to TNS severity (i.e., score < 5 vs ≥ 5; see table 2); a positive result was obtained in the patients with a milder neurotoxicity, but not in the most severe cases. The statistical results were unchanged when, besides the examination level, a patient level comparison was also performed (i.e., using only one visit in those patients examined more than once, data not shown).

Regarding the individual TNS items, we found that the VDT examination did not behave in a linear way when compared with TNS (figure, A), despite a significant correlation (r = 0.615, p < 0.001). In fact, for TNS scores below 12, VDT values ranged between 0 and 6 (normal value ≤ 2), but most of the values were normal or showed minimal changes against a wide range of pathologic TNS scores. When the TNS increased over 12, the VDT slope changed markedly, thus making the clinical association between VDT and CIPN rather unreliable. The sural nerve SAP and peroneal nerve CMAP results showed a significant correlation with the severity of CIPN (r = −0.405, p < 0.001 for SAP, r = −0.305, p = 0.032 for CMAP, figure, B and C), but also in these cases the presence of clinical signs of CIPN was frequently associated with amplitudes that were normal.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Correlation studies between Total Neuropathy Score (TNS) and vibration detection threshold (VDT) (a) and sural nerve (SAP) (b) and common peroneal (CMAP) (c) in the studied population (95% confidence limits).

When the TNSr was compared to the oncologic toxicity scales, the results could be superimposed on those obtained with the extended (i.e., complete) TNS (see table 2).

Discussion.

The issue as to which is the best method for investigating and reporting the type and severity of CIPN remains unsolved.

The aims of the study were to determine in a homogeneous series whether the TNS composite score7-9⇓⇓ might be a reliable method for assessing the presence and severity of CIPN and to ascertain whether its results can be correlated with clinically based toxicity scales commonly used by oncologists.4-6⇓⇓

The TNS is a composite scale with a much larger range of values (from 0 to 40) than the oncologic toxicity scales (ranging from 0 to 4 grades). It is conceivable that TNS would be more appropriate for investigating and accurately reporting the severity of CIPN than a five-grade scale. However, the correlation between TNS and the neurotoxicity oncologic scales has never been assessed. Therefore, the use and interpretation of TNS results may pose problems for clinical oncologists.

TNS reliably reflects the overall changes corresponding to grades 0 to 2 of the oncologic scales, a severity that represents the range of neurotoxicity most commonly encountered in the treatment of patients with cancer. Moreover, the more extended range of values of TNS clearly allows a wider possibility of scoring the severity of CIPN. In fact, a ceiling effect is present with the common toxicity scales and in our study most patients are scored with the same score value of 2 when CIPN becomes clinically relevant. TNS values, on the contrary, change more gradually, thus allowing minor changes in CIPN symptoms and signs to be recognized more efficiently. This different behavior explains the absence of a statistically significant correlation with the common toxicity scales when TNS score is ≥ 5. However, TNS requires time (about 1 hour), technical facilities, and specific expertise—all requirements that make it more suitable for research purposes than for routine clinical practice. Therefore, we determined whether a reduced TNS (TNSr, range 0 to 28) may be more easily adopted for assessing the severity of CIPN in routine clinical activity.

TNSr allows the same correlation to be obtained with the oncologic toxicity scales as that observed with the extended TNS version. The three items excluded by the TNS in order to obtain the TNSr were selected according to 1) the predictable likelihood of their occurrence with the treatment schedule adopted in this study and 2) the general availability of the device used for the instrumental examination.

We confirmed that reliable and reproducible grading of CIPN can be obtained with TNS, but TNSr can be used as an effective and faster alternative. Moreover, it is extremely important in order to compare our neurologically oriented results with the oncologic data that TNS and TNSr results are clearly correlated with the clinically relevant results of NCI-CTC 2.0, Ajani, and ECOG scales. Finally, a flexible use of the TNS items makes it possible to tailor the scale to the predictable toxicity profile of each chemotherapy schedule.