Diagnostic tools should be used for the diagnosis of chemotherapy induced peripheral neuropathy in breast cancer patients receiving taxanes

Abstract Background Though the incidence, characteristics, and pathogenesis of chemotherapy induced peripheral neuropathy (CIPN) by taxane based chemotherapy were extensively studied, diagnostic guidelines extent only recently. Aim To observationally investigate whether specific tests can be used to predict and monitor CIPN severity. Methods Fourteen female breast cancer patients receiving paclitaxel or docetaxel were evaluated using the McGill Pain Questionnaire (MPQ), National Cancer Institute Common Toxicity Criteria (NCI‐CTC) grading, clinical total neuropathy score (TNSc), quantitative sensory testing (QST) of pressure pain threshold (PPT), and numeric rating scale (NRS) scores and stocking and glove distribution testing (SGDT), at the start (T0), midst (T1), and end (T2) of their treatment and after 3 months (T3). Results At T3, patients scored NCI‐CTC neuropathy grade 1 (14.3%), 2 (64.3%), and 3 (14.3%) respectively. Fifty percentage scored at least grade 1 at T0, with complaints not caused by CIPN. Pain, if present, was denominated “tingling” and “cold” in the MPQ. Median TNSc score increased from T0 (2.43) to T1 (4.71) to T2 (5.50) to T3 (5.57), as did pinprick and cold sensation disturbances in SGDT. PPT and associated NRS remained unchanged. TNSc and SGDT at T1 could not predict the NCI‐CTC grade at T3. Conclusion NCI‐CTC, TNSc, and stocking and glove distribution testing can be used in the early diagnosis and monitoring of CIPN, with false‐positive findings at baseline. Final NCI‐CTC grades could not be predicted.


| INTRODUCTION
Yearly, 1.7 million women are diagnosed with breast cancer worldwide.
Treatment can comprise (a combination of) surgery, radiotherapy, chemotherapy, monoclonal antibody therapy, and hormone blocking therapy, depending on disease stage and specific gene expression. 1 To improve survival, anthracycline and cyclophosphamide therapy is often combined with a taxane. 2 Chemotherapy induced polyneuropathy (CIPN) is a commonly reported side effect (incidence up to 68.1%), prompting dose reduction or discontinuation, thus worsening outcome. 3,4 While a good target for antineoplastic drugs, microtubules (intracellular structures involved in movements in eukaryotic cells and in mitosis) play a role in the pathogenesis of CIPN by taxanes (and vinca alkaloids).
Usually they are in a state of dynamic instability, a process dependent on (de)polymerization during which the ends of microtubules switch between phases of growth and shortening. 5 Taxanes prevent the depolarization of microtubules, thus impeding normal dynamic instability essential for cell division in malignant cells (hence: "microtubule stabilizing agents"), and microtubule-based axonal transport.
They also cause macrophage activation in dorsal root ganglia and peripheral nerves and microglial activation within the spinal cord, resulting in deficits in signal transduction, progressing from distal to proximal. [6][7][8] Since the introduction of taxanes, the incidence, and characteristics of CIPN have been extensively described, using different scales for grading (e.g., the World Health Organization [WHO], Eastern

Cooperative Oncology Group [ECOG], Ajani and National Cancer
Institute Common Terminology Criteria [NCI-CTC] scales). All grade neuropathic symptoms between 0 (no symptoms) and 4 (severe symptoms, loss of function), grade 3 usually triggering chemotherapeutic agent dose correction. [9][10][11][12] Despite different grading systems, diagnostic and screening guidelines are only recently build. 13,14 In this prospective and observational feasibility study we evaluate clinical tools for screening, early diagnosis and evaluation of CIPN in breast cancer, evaluating pitfalls and practicality of their application by comparing results for different tests and points in time. For broad application, tools should be easy to use, widely available and validated. Therefore, we selected questionnaires and neurological tests meeting these criteria: McGill Pain Questionnaire Dutch Language Version (MPQ-DLV), NCI-CTC, clinical version of the total neuropathy score (TNSc), quantitative sensory testing (QST) of the pressure pain threshold (PPT), and stocking and glove distribution testing (SGD testing).

| Study design and setting
For this prospective observational cohort study, patients were recruited and assessed in the Alexander Monro breast cancer institute's outpatient clinic in Bilthoven between February 2017 and October 2017.The study protocol was reviewed and approved by the regional ethics committee (CMO Regio Arnhem-Nijmegen). Written informed consent was obtained from all patients for participation in the study and publication of anonymous data. Standard care was provided. The STROBE statement checklist for observational studies (cohort) was followed if applicable. The data that support the findings of this study are available from the corresponding author upon reasonable request. Of 21 screened women, we included 16 female patients diagnosed   with of breast cancer scheduled for treatment with taxane based chemotherapy, aged 18 years or older. Patients with a history of chronic   pain, pre-existing neuropathy, a history with risk factors for poly-neuropathy (e.g., alcohol abuse, diabetes mellitus, vitamin B12 deficiency) or not able or willing to give informed consent, were excluded.

| Measurements
Fourteen out of sixteen included subjects were interviewed and neurologically assessed four times during chemotherapy cycle: at the start (T0), midst (T1), and end (T2) and after 3 months (T3). At baseline, age, ethnicity, medical history, height, weight, current medication, tumor characteristics, and treatment characteristics were recorded. At each assessment point patients filled out the MPQ-DLV, the NCI-CTC grade and TNSc were determined, and QST of PPT and SGD testing took place. A NCI-CTC grade of CIPN at T3 was considered the final outcome of severity.
For the further tests, their predictive value at earlier stages for the NCI-CTC grade at T3, and thus long-term severity, was investigated.

| National Cancer Institute Common Toxicity Criteria
The Indication for Common Toxicity Criteria Grading of Peripheral Neuropathy Questionnaire (ICPNQ) was used to determine the severity of CIPN on the NCI-CTC scale, grading 1 (sensory changes or pain), 2 (any autonomic symptoms or subjective motor weakness), 3 (symptoms lead to ADL impairment), or 4 (ADL dependency). 15 includes evaluation of sensory symptoms, motor symptoms, autonomic symptoms, pinprick sensibility, vibration sensibility, strength, tendon reflexes, scoring 0-4 on 7 items. It has a positive correlation with the NCI-CTC grading system and an even higher sensitivity to CIPN changes. 18 The full TNS also scores vibration sensation (QST of vibratory detection threshold) and nerve conduction studies (sural and peroneal amplitudes), thus scoring 0-4 on 10 items.
Severity of sensory, motor, and autonomic symptoms were evaluated by asking patients for localization of symptoms, extent of limitation and number of symptoms, respectively. Pinprick sensibility was tested in the upper and lower extremities. Vibration sense was tested by placing a vibrating tuning fork (128 Hz according to Hartmann, Rudolf Riester GmbH, Jungingen, Germany) on bony prominences of the upper (distal interphalangeal and proximal interphalangeal joints of the index finger, ulnar styloid process, olecranon) and lower extremities (first metatarsophalangeal joint, lateral malleolus, patella), bilaterally. If vibration sense in the most distal testing point was not disturbed, more proximal points were not tested and assumed to be normal (for TNSc: reduced in fingers/toes, up to wrist/ankle or up to/ above elbow/knee). 19 Strength of extension and flexion of the hands and feet was evaluated according to the Medical Research Council scale. Last, the deep tendon reflex of the left and right Achilles tendons was tested. If the Achilles tendon reflex was reduced or absent, the patella tendon reflex was tested as well.

| Quantitative sensory testing
Patients with neuropathic pain often develop mechanical hyperalgesia. 20 By testing the PPT in our patients, we explore if the same is true for patients with CIPN. PPT was measured with the Wagner Force Ten™ FDX digital force gage; a handheld pressure algometer. The pressure was exerted on the left and right musculus trapezius pars medialis (4 cm paravertebral of vertebrae T3), musculus rectus femoris (10 cm proximal to patella), thenar eminence (on the middle of the muscle) and musculus abductor hallucis (on the middle of the muscle). Patients were instructed to tell the researcher to stop applying pressure if it became too uncomfortable or painful. The applied pressure in kilopascal (kPa) was then recorded. Patients were asked to score pain according to the Numeric Rating Scale (NRS).

| SGD testing
Sensory CIPN symptoms in patients treated with taxanes often emerge in a stocking and glove pattern. 20,21 To assess sensory changes, cold sensibility was measured delivering stimuli with an ice cold Tip-Therm and pinprick sensibility was measured delivering stimuli with a BD™ 18 G Blunt Fill Needle. Stimuli were given every 3 cm from the cubital fossa to the top of the third digit and from the patella to the top of the first (big) toe on the left and right side. Patients were instructed to indicate changes in sensation (hypo-or hyperalgesia) along the testing points. If present, the distance (cm) between the beginning and the end (often being the tip of digit III or the great toe) of the area with changed sensibility was recorded.

| RESULTS
Of 16 enrolled patients, 2 patients were excluded at T2 and T3 respectively because of missed appointments and patient withdrawal.
Patient characteristics are shown in Table 1.
Dose reduction was required in six subjects due to toxicity, including severe neuropathy in two cases. In one of these two patients, taxane therapy was reduced by two cycles due to the neuropathy symptoms. We ran two sets of analysis of data-first to compare the outcomes at T0, T1, and T2 (complete study population) and then a second analysis to compare the outcomes of the patients tested at T3 with their previous outcomes (within subject comparison).

| National Cancer Institute Common Toxicity Criteria
CIPN grade distribution at each measuring time are shown in Figure 1.
According to the ICPNQ questionnaire, nine patients already had At T3 all but one patient suffered from neuropathy.

| NCI-CTC versus TNSc
The correlation between NCI-CTC grade and TNSc score was 31.8% (p = .017) for all 56 measurements. Graphical representation of the different measurements (14 patients at 4 time points) in Table 2.

| McGill Pain Questionnaire Dutch Language Version
At T0 three patients indicated to experience pain: two described it as "tingling" (left hand and right foot respectively) and one as "pulling" (left hand). At T1 again three patients had pain; the two patients with the "tingling" feeling still among them, although one now described the pain as "cold." The patient with the "pulling" pain at T0 had no pain at T1. One patient had new onset of pain in the feet, described as "mild" and "pricking." At T2 seven patients reported pain, mainly described as "tingling" and "tight." At T3 5 out of the 11 patients still had pain, predominantly described as "tingling" and "cold."

| Quantitative sensory testing
For all measurements in one specific subject PPT could not be found,  Table 3.

| SGD of pinprick and cold testing
Paresthesia usually starts at the most distal nerve endings (fingers and toes) and progresses proximally. 4  F I G U R E 2 Mean length (cm) of the affected areas as assessed by stocking and glove distribution (SGD) testing for pinprick and cold testing ("pinprick" and "cold" respectively) of the upper and lower extremities ("upper" and "lower" respectively) found areas of hypo-or hyperalgesia for pinprick or cold sensation, but not in a stocking and glove distribution (e.g., 10 cm hypoalgesia in the forearm when testing pin prick sensation, but no sensory disturbances in de hand or fingers). The analyzed length of the affected areas over time are shown in Figure 2.
Friedman testing in the presence of multiple outliers and abnormal distribution, presented significant differences in outcomes for pinprick in upper (χ 2 = 9.14; p < .05) and lower extremities (χ 2 = 21.29; p < .01) and cold sensation in upper (χ 2 = 17.08, p < .01) and lower extremities (χ 2 = 17.04; p < .01) at T0, T1, and T2 for the complete study population. Results of post-hoc Wilcoxon signed-rank testing are presented in Table 4. also received carboplatin treatment, emerging complaints of CIPN did not occur significantly more than in the other subjects. They were therefore included in the further analysis regarding the use of diagnostic tools. Theoretically however, they could influence the different scores, by the mechanisms and specific complaints to which they cause CIPN.

| DISCUSSION
Despite the small study population, this study addresses the possibilities as well as the pitfalls of different tests used in the (early) recognition and the monitoring of the progression of CIPN related complaints. For women with breast cancer receiving taxane based chemotherapy, NCI-CTC and TNSc scales, as well as the "stocking and gloves" distribution of pinprick and cold testing abnormalities, but not (threshold) pinprick testing and associated NRS scores, are useful for determining the progression of CIPN symptoms. However, confounders influence (from baseline onwards) the definitive score. In conclusion, this feasibility study shows that diagnostic tools should be used for the diagnosis of CIPN. A definitive approach cannot be formulated, due to the influence of confounders and lack of research addressing different cancer and chemotherapy types in relation to the wide variety of available tools.