Patient-reported outcomes in chemotherapy-induced peripheral neuropathy: a review

The Patient Neurotoxicity Questionnaire (PNQ) has two versions: one two-item version to be administered for studies of taxanes, cisplatin, and carboplatin and a two-item version to be administered for studies of oxaliplatin.² All the measures include symptoms such as tingling, numbness, pain, and/or cold perception in the hands or feet, although there is variability in item wording and whether or not selected symptoms are shown together with an “or” conjunction (Table 3). With the exception of the PNQ, which combines symptoms into two items, the measures generally include a “laundry list” of symptom items that also are separated by whether or not they occur in the hands versus the feet.

There is some indication that the impact of CIPN is multidimensional. For example, in an analysis of PNS data, a principal component factor analysis yielded a two-factor solution comprising “hand neuropathy” and “foot neuropathy.”²⁶ In a psychometric analysis of the Functional Assessment of Cancer Therapy/Gynecologic Oncology Group—Neurotoxicity (FACT/GOG-Ntx) scale, Huang and colleagues²⁷ divided the items into four subscales: sensory, motor, hearing, and dysfunction. The four items of the Ntx scale related to sensory neuropathy contributed the most to CIPN-related treatment differences. Nevertheless, most of the measures simply combine all items into an overall score, even those that comprise both symptom and activity limitation items.

The PRO measures vary with respect to development status. Except for the PNQ, the content of all the measures was informed in some way by patient input (this information is unknown for the CIPNS-32 and APN). However, patient assessment of content validity through cognitive debriefing interviews has only been reported for the PNS and CIPN20. When assessed for the measures, internal consistency reliability (Cronbach's alpha) is greater than 0.70, which is acceptable for making group comparisons. None of the measures reports test–retest reliability estimates.²²

The CIPN20 is being developed following the steps recommended in the Food and Drug Administration (FDA) PRO Guidance, and it currently is being field-tested in a large, randomized clinical trial.^{[1] and [28]} The FDA PRO Guidance recommends conducting one-on-one interviews and/or focus groups with patients for item generation, performing cognitive interviews to obtain feedback on the measure from patients, and performing a psychometric validation study when developing a new instrument. A personal communication with the authors of the CIPNS-32 revealed that no additional information on psychometric properties was available. The PNQ has been used to assess the impact of chemotherapy among adjuvant breast cancer patients.²⁹ In this study, the PNQ appeared to detect more severe grades of sensory neuropathy compared to the NCI-CTC.²⁹

Among the neurotoxicity measures, the FACT neurotoxicity scales have been most widely used in randomized controlled trials. The scales evaluate the impact of treatment on CIPN as an outcome on three specific areas of functioning—sensory, motor, and auditory—and have frequently demonstrated sensitivity to treatment effects.^{[7], [30], [31], [32], [33] and [34]} For example, Land and colleagues³⁴ used the FACT/GOG-Ntx-12 scale in colon cancer patients treated with fluorouracil, leucovorin, fluorouracil + oxaliplatin (OF) or leucovorin + oxaliplatin (OL). The mean scores for patient-reported neurotoxicity on the FACT-Ntx scale were significantly worse in the oxaliplatin arm throughout the 18-month period of the study (P < 0.001). Time to neuropathy resolution was significantly longer in the oxaliplatin group and continued beyond 2 years in more than 10% of patients.³⁴

Moreover, neurotoxicity-specific measures like the FACT Ntx scales have been shown to be more sensitive than laboratory- and physician-based measures in detecting CIPN.^{[3], [10] and [35]} For example, in a study of amisfostine for treatment of CIPN, the FACT/GOG-Ntx scale was more sensitive than the vibration perception threshold and NCI-CTC in detecting CIPN.³⁵ In another randomized controlled trial comparing whole abdominal irradiation and doxorubicin–cisplatin in advanced endometrial cancer, the APN was found to be more sensitive than the FACT-G in detecting changes related to CIPN.²⁴

Customized items to assess CIPN have also shown sensitivity to treatment effects and the ability to capture the occurrence of CIPN over the long term. For example, Bezjak et al.³⁶ used one CIPN-specific item about “sensation in fingers or toes” in addition to the EORTC QLQ-30 to evaluate the quality-of-life impact of a cisplatin-based regimen among ovarian cancer patients. The study found significant treatment group differences in the CIPN item at every follow-up, although differences in the global or physical, emotional, social, and role function domains of the EORTC-QLQ-30 were not significant. In addition, the brief SCIN was found to be particularly useful for capturing lingering long-term toxicities of cisplatin and bleomycin among testicular cancer patients.^{[37] and [38]}

Discussion

The impact of CIPN is unanticipated by patients and underestimated by clinicians. CIPN symptoms often linger well after termination of treatment, sometimes leading to prolonged functional impairments and suboptimal quality of life. Key physical impacts of CIPN include symptoms such as tingling, numbness, and pain in the hands and feet and activity limitations such as difficulties with walking, cooking, dressing, bathing, hobbies, and driving. Patients, often reluctant to discontinue chemotherapy, live with CIPN symptoms and express helplessness, anxiety, and frustration for not being able to manage CIPN-related disabilities and disappointment that clinicians often fail to appreciate its impact on everyday life. The disconnect between physicians' and patients' perceptions of CIPN was noted in several studies and is attributable, in part, to the lack of comprehensive and easily interpretable tools to measure CIPN.

Currently, there is no single comprehensive PRO scale to evaluate CIPN symptoms, severity, functional impairment, and psychosocial impacts, although several neurotoxicity PRO scales have demonstrated acceptable validity, internal consistency, and construct validity. Therefore, additional content validity assessment may be necessary if these measures are to be used to obtain labeling claims in the United States.

Of the neurotoxicity PRO scales, the FACT/GOG-Ntx scales have been used most often in randomized clinical trials and have demonstrated acceptable reliability, validity, and responsiveness.^{[32], [33] and [38]} The Ntx scales were developed as modules for administration with their core measure, FACT-G. The EORTC-QLQ CIPN20 is under development and, given that it specifically focuses on peripheral neuropathy, may be more sensitive in measuring CIPN-related functional impairment compared to FACT scales.

A scale capturing emotional distress would be a useful addition to several of the existing CIPN measures, enabling a thorough evaluation of CIPN. Such a scale may be sensitive to treatment group differences in a clinical study. As observed by Bruner et al.,²⁵ the APN, which measures bothersomeness, was more sensitive than the FACT-G in detecting changes in CIPN over time. The CIPNS-32 also inquires about bothersomeness of impacts, but unfortunately development information on this scale is not published.

Ultimately, the selection of an appropriate CIPN PRO measure will depend on the objective of the study and the future uses for the data; for example, if one is interested in assessing only the occurrence of CIPN, only a few items likely are required. If, on the other hand, one is interested in a comprehensive assessment of CIPN impacts, including assessment of functional and emotional impacts, one would need to develop a new measure or use a combination of available PRO scales. Dunlap and Paice³⁹ have suggested that standardized instruments that evaluate peripheral neuropathy in those receiving chemotherapy should be multidimensional, addressing the nature, intensity, and time course of symptoms, as well as their effect on quality of life.

Key domains that likely should be captured within a comprehensive PRO measure of CIPN impacts include hand and foot symptoms, central nervous system symptoms, daily physical functioning/activities, emotional impacts such as frustration and anxiety, and work productivity.

However, to ensure that all important concepts are covered, it would be useful to conduct focus groups or one-on-one interviews with patients experiencing CIPN. Although the qualitative and preference studies report on patient experience of CIPN, additional research among CIPN patients would help us to understand the patient experience fully. Studies that evaluate the relative importance of functional impairments and patient strategies used to cope with these impairments are necessary. Such studies will not only help to identify constructs that are important to patients but can also be used to inform a conceptual model. Similarly, patient preference research that sheds light on the relative preferences of patients with previous experience of CIPN versus those suffering from CIPN for the first time or the disutility associated with CIPN and other side effects would help clinicians to optimize cancer treatment by incorporating the patient's view in clinical decision making.

Conclusion

In the larger context of treating carcinomas, CIPN may seem like a self-limiting and acceptable cost considering the benefits of chemotherapy. However, its impact on patients' HRQL and sometimes on overall treatment outcomes can be significant. The lack of effective pharmacological options makes CIPN symptom management challenging, and teaching patients effective coping skills becomes a large part of CIPN treatment. Often, physicians do not ask patients about neurological symptoms unless patients complain about them, and patients may also have difficulty understanding their own symptoms as attributable to CIPN. It is therefore essential to educate patients and provide them with language to express CIPN symptoms as well as teach coping mechanisms to effectively manage CIPN.

Physician-based measures such as NCI-CTC and laboratory-based measures such as nerve conduction studies often underestimate the severity of CIPN symptoms.^{[3], [37] and [39]} Researchers note that the CIPN incidence and severity data should be collected from patient self-report questionnaires as studies show that physicians underreport the impact of CIPN on patients.^{[3], [37], [39] and [40]} As newer chemotherapies with more neurotoxicities emerge, the evaluation of neuropathy is likely to gain greater importance in patient-reported outcomes.

This review underscores the need to use neurotoxicity-specific PRO measures to detect meaningful differences in CIPN HRQL impact between treatments and to obtain patient self-reports of CIPN symptoms to accurately account for the influence of CIPN on HRQL. Outcomes of CIPN are not measured in a uniform fashion. For consistent measurement of patient-reported outcomes of CIPN, first clinicians and researchers need to recognize that CIPN has a significant impact on patient's HRQL, use comprehensive scales with adequate patient input to evaluate those outcomes, and teach patients effective coping mechanisms