Chemotherapy-induced peripheral neurotoxicity – outcome measures: the issue

KEYWORDS:

• Chemotherapy induced peripheral neurotoxicity
• outcome measure
• quality of life
• translational medicine
• neurophysiology

1. Introduction

Chemotherapy-induced peripheral neurotoxicity (CIPN) has become a hot topic in oncological daily practice. It is a common condition, related to chemotherapy regimen used to treat breast, colorectal, and lung cancer; platinum drugs, taxanes, vinca alkaloids, proteasome inhibitors, epothilones, and thalidomide are mainly implicated in the development of CIPN [1]. In recent times, cancer patients have become a long-surviving population, thanks to advancement in treatment with these drugs and amelioration in cancer diagnosis. Mortality, adjusted for age, has decreased both in pediatric and in adult populations between 1950 and 2010 [2]. Therefore, chronic or late toxicities, such as CIPN, have become an emergent issue.

CIPN is a sensory, length-dependent neuropathy/neuronopathy. In rare cases, motor or autonomic impairment or cranial nerve involvement has been observed. Different drugs present a different clinical picture [1]; however, some features are quite consistent. Sensory alterations can be divided into two main categories. ‘Negative’ signs/symptoms: impairment in touch, pin, and vibration perception. If there is a consistent loss of large fiber modalities, the patient can develop disability, even though motor function is preserved. Sensory ataxia can determine imbalance and falls and difficulty in manipulating objects; this is known as loss of ‘composite function.’ ‘Positive’ symptoms are mainly related to damage of small fibers: paresthesia/dysesthesia and neuropathic pain. Alterations develop initially at limb extremities after which they have a distal-to-proximal progression. The damage is usually dose dependent. In some cases, in particular with platinum drugs, neuropathy can worsen for a few months after chemotherapy termination, the so-called ‘coasting phenomenon’ [3].

All described manifestations have a negative impact on quality of life (QoL) causing pain or disability. So far, a preventive or a curative strategy has not yet been found as efficacious, as accurately pointed out in a meta-analysis recently published by the American Society for Clinical Oncology [4]: only duloxetine has been recognized as moderately efficacious as a symptomatic option [4,5]. Some methodological issues at bench and at bedside can explain this lack.

2. Clinimetric issues

There are still no definite epidemiological data on CIPN since a gold standard in its assessment has not yet been established. Prevalence/incidence data were collected with different outcome measure(s); so, it is quite difficult to analyze literature to obtain comparable data.

When approaching peripheral nervous system dysfunctions, clinical examination and quantitative methods, such as nerve conduction study (NCS), are commonly used in clinical practice. In the research setting, many different clinical scores were developed to rate CIPN, taking into account signs and symptoms. Despite great effort has been done in the past two decades, so far, the ideal instrument(s) to be applied has not yet been defined [6]. Ideally, the gold standard outcome measure(s) should be reliable and valid, but also responsive to modification of neurological examination/symptoms of the patient. There are two main categories of tools to grade CIPN: physician-based outcome measures (i.e. clinical scale) and patient-reported outcome measures (PROs, i.e. questionnaires).

The first ones are mainly based on the clinical examination and the physician point of view; the second ones, instead, are direct information given by patients themselves.

Clinical scales were initially developed in an oncological setting as part of the general adverse event detection; one of the main applied scale, nowadays, the National Cancer Institute Common Toxicity Criteria (NCI-CTC) [7], in fact, was developed for this purpose. However, despite being a physician-rated outcome measure, it is not based on a formal neurological examination; other tools were then developed to overcome this flaw [6]. Another commonly applied clinical scale was, in fact, developed: The Total Neuropathy Score – clinical (TNSc) scale, which is mainly based on the neurological examination. TNSc has been proposed as a potentially useful tool to be implemented in everyday practice and in clinical trials [6].

In the last two decades, given the importance of QoL, great effort was also extended to develop valid and useful PROs. Some of the most applied ones which specifically detect neuropathy symptoms are: Functional Assessment of Cancer Therapy/Gynecologic Oncology Group-neurotoxicity, Functional Assessment of Cancer Therapy-Taxane, peripheral neuropathy scale, and oxaliplatin questionnaire [6]. They are useful tools that allowed a better insight on patient QoL. They were designed to detect symptoms that are commonly known to burden patients affected by CIPN. Despite having been a great advancement in CIPN evaluation, none of them reached the consensus to be elected as THE gold standard. Moreover, there are still many PROs that are differently employed in different clinical trials; this, again, makes difficult to compare data among different studies. Ideally, a unique and comprehensive PRO for CIPN patients could enable to gain even more informative data, which could be promptly compared among different studies.

On the basis of these lacks, a consistent methodological study was designed, CI-PeriNomS study. This was an observational clinical study. Its aim was to better investigate physician-based outcome measure(s) and PROs, in a well-selected population of CIPN patients. First validity and reliability findings have been obtained for the selected outcome measures; initial answers to issues addressed so far were, in fact, obtained [8]. In particular, NCI-CTC and TNSc were employed; it was demonstrated that TNSc has a good clinimetric profile since it showed good reliability. TNSc cannot be defined yet as ‘the’ gold standard in CIPN, but it can be suggested that TNSc should be even more implemented in CIPN evaluation, both for clinical and for research purposes. The CI-PeriNomS study gave also highly informative data about PROs. It was demonstrated that PROs gave additional information to those obtained by the physician; PROs and physician-rated parameters gave different information that were complementary. Thus, it can be suggested that a comprehensive CIPN assessment cannot be performed if both physician-based and PROs are not applied together [9]. CI-PeriNomS did not discovered the ideal tool(s) to be universally suggested in CIPN evaluation, but precious information was gained. Moreover, the same research group is performing a subsequent study, CI-PeriNomS-II, which will give an insight also on responsiveness of the tools employed; the study is expected to be closing in late 2017.

3. Preclinical issues

Among reasons for lack of a valuable CIPN treatment, there is also insufficient knowledge of the pathogenesis of cancer treatment-related neurotoxicity. To achieve a better understanding of chemotherapy-induced neurotoxicity, accurate preclinical studies might be fundamental, but the translation of results into clinical trials is sometimes difficult. It is important to carry on preclinical research on a refined and well-characterized model. Several in vivo rat models have been established over the past 20 years to reproduce CIPN phenomena and to study the underlying mechanisms responsible for peripheral nervous system damage [10–13]. Rat models are currently being replaced by mouse ones. Neoplasms cannot be easily reproduced in immunocompetent rats. Mice models are, instead, nowadays widely employed in many oncological preclinical studies to establish the activity of chemotherapy. Their application in CIPN is still as not as wide, even though in the last few years, models in this specie have started to be proposed [14]. In immunodeficient mice, a neoplasm can be inoculated reproducing the condition of patients which will be treated for effectiveness against the tumor; thus, when testing a compound for CIPN prevention, safety and interaction with chemotherapy can be also tested [15]. At the moment, the aim of a good model in CIPN is to reach a real ‘chronic’ neuropathy induction and the application of formal assessment methods, in order to fully characterize neuropathy. But a way to better translate preclinical data into clinical trials is still awaited. Neurophysiology could be the answer since there are many techniques that could be implemented. NCS and microneurography (MN) could be useful to study a condition as CIPN is. NCS are routinely performed in patients to diagnose neuropathy. NCS are able to detect mainly large fiber damage; alterations in these fibers are related to sensory loss and sensory ataxia. MN instead could make possible also small fiber damage detection; damage of these fibers is the anatomical correlate of neuropathic pain. MN could then give valuable information about pain which is not easily detected and graded in animals. MN is also performed in the clinical setting, even if not routinely. So, combining NCS and MN could allow to detect in animals all alterations that can be compared with the full clinical spectrum in patients.

4. Expert opinion

CIPN is a challenging issue that would greatly benefit from a high translational research approach. So far, in fact, CIPN cure has been ineffective due to unmet needs both at bench and at bedside. In the next few years, investigation in this regard will be fundamental to find a cure for a condition that is detrimental for QoL of a long-surviving population. At the bedside, a clinimetric approach is to be implemented. Ideally, a definite and universally employed outcome measure should be sought for. This outcome measure should be a composite one, combing both physician-rated items and patient-rated ones. If all clinical trials, in future, will be performed with the same valid tool, data will be more solid, and different studies could be better compared.

At the bench side, innovative methods to promptly translate data to bedside are a key issue. In this regard, neurophysiological testing is the ideal candidate for this objective. Neurophysiology can detect and grade nerve damage with the same approach applied in humans. NCS is appropriate to detect, in particular, large fiber damage, and MN could be useful to detect small fiber ones. Combing these techniques could make animal model assessment rather comprehensive. Thus, neuroprotectant agents, in future, could be tested both in animals and in humans with the same tool. Neurophysiology could become the link between bench and bedside; at bench side, it is possible to easily perform histopathology, but it is difficult to manage complex items such as neuropathic pain is. Instead, at bedside, we can employ PROs to gain information about QoL, but histopathology is difficult to be implemented. Thus, neurophysiological determinations could make possible to indirectly correlate histopathology in animals with PROs as rated by patients. This correlation is even more interesting in the pharmacogenomic and personalized medicine era. Pathophysiological studies could be done in animals. Pharmacogenomics studies are better suitable to be performed at a clinical level. If the two ‘side of the coin’ could have a similar tool for neuropathy determination, data could be better and easily translated from preclinical hypothesis to clinical evidences.

Declaration of interest

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Additional information

Funding

This paper was not funded.