Assessment of neuropathic pain following cancer treatment

Neuropathic cancer pain (NCP) is prevalent affecting up to 58% of those with persistent pain following cancer treatment. Neuropathic pain can develop from malignancy, after neural tissue insult during surgery and/or exposure to radiation or neurotoxic agents used as part of cancer treatment regimens. Pain following cancer treatment is commonly under‐treated and one barrier identified is poor recognition of pain and inadequate assessment. Recognition of the presence of NCP is important to inform pain management, which is challenging to treat and warrants the use of specific treatments to target neuropathic mechanisms. In this review, approaches for screening and classifying NCP are described. These include screening questionnaires and the application of the updated neuropathic pain grading system in a cancer context. The evidence from neuropathic pain related assessments in cancer populations is provided and highlighted under different neuropathic pain grades. Recommendations for assessment in practice are provided.


| INTRODUCTION
Improvements in oncology mean extended periods of cancer survivorship; however, pain is a common impediment to optimal quality of life in cancer survivorship (Bennett et al., 2019).Cancer-related pain may be caused by the primary cancer itself or metastases (cancer pain) or its treatment (post cancer treatment pain) (Bennett et al., 2019).Neuropathic pain (NeP) is defined as pain caused by a lesion or disease of the somatosensory nervous system (Jensen et al., 2011).Neuropathic cancer pain (NCP) can be related to the primary cancer or metastases or may be a consequence of post-cancer medicine, radiotherapy or surgery (Bennett et al., 2019).
The prevalence of cancer-related pain is high with pooled data (n = 122 studies) indicating rates of 39% after curative treatment; 55% during anticancer treatment; and 66% in advanced, metastatic, or terminal disease (van den Beuken-van Everdingen, et al. 2016).The pooled prevalence of NCP in the general cancer population has been reported to be up to 44% among adult cancer populations, (Roberto, et al. 2016) while pooled data on specific populations for example, early breast cancer, indicate a prevalence up to 58% of those with pain (Ilhan et al., 2017).Unfortunately pain during and following cancer treatment is frequently under-treated (Kojima, et al. 2012), one reason being inadequate assessment (Deandrea, et al. 2008).Further challenging classification of NCP is the finding that up to two-thirds of those with NCP present with mixed pain that is, a combination of NeP plus other pain types (Bennett et al., 2012;Leysen et al., 2019).Distinguishing NCP is important clinically and for research purposes as those with NCP demonstrate higher pain intensity, analgesic consumption and reduced quality of life compared with other cancer survivors (Bennett et al., 2012;Fainsinger et al., 2010;Garz on-Rodríguez et al., 2013;Rayment et al., 2013) and hence, warrant specialist medical and allied health care.
In 2011, the NeP special interest group (NeuPSIG) of the International Association for the Study of Pain (IASP) endorsed a grading classification system for NeP (Haanpää et al., 2011), which was updated in 2016 (Finnerup et al., 2016).In cancer populations, there has been a dominance of classification of NeP based on using screening questionnaires for NeP although some recent studies have compared the use of screening questionnaires with clinician-derived assessment using the NeP grading system (Mulvey et al., 2017;Shkodra et al., 2021).While screening questionnaires are clinically practical and convenient and reasonably accurate (Mulvey et al., 2017), they may under-estimate prevalence of NeP, which has been found to be higher when assessed by clinical impression or classification (Shkodra et al., 2021).Further, over the last decade, a large body of research has emerged investigating various pain features, distributions, psychophysical and objective measures related to possible NeP in cancer populations.Outlines of clinically relevant processes and important considerations for applying the NeP grading system for NCP have previously been published (Boland et al., 2015;Edwards et al., 2019;Mulvey et al., 2014).However, these previous articles did not integrate specific research findings from cancer populations, which have emerged over recent years.This is warranted to enhance nuanced application of the NeP grading system and interpretation of assessment findings based on the evidencebase specific to cancer populations.
Therefore, the purpose of this narrative review was to synthesize recent evidence on features of NCP across cancer populations, present these findings in the context of the updated NeP grading system, and to highlight common presentations to refine recommendations for assessment and grading NCP and interpretation of findings.

| METHODS
A narrative review in line with the scale for the assessment of narrative review articles (SANRA) was conducted (Baethge et al., 2019).The primary aim was to synthesize recent literature on the assessment of NCP using the NeP grading system (Finnerup et al., 2016;Haanpää et al., 2011) which has also been recommended for use in cancer pain patients (Shkodra et al., 2021).The grading system focuses on the distinction between neuropathic and non-NeP, and identifies four levels of certainty, namely unlikely, possible, probable, and definite NeP.The flowchart of the updated grading system, with its components can be consulted in Figure 1.
F I G U R E 1 Flowchart of the updated grading system for neuropathic pain.From (Finnerup, et al. 2016)

| INCLUSION/EXCLUSION CRITERIA
Studies were included if they met the following criteria: original research and systematic reviews published since 2008 after the NeP grading system was first published (Treede et al., 2008); assessed clinical features, psychophysical tests or objective assessments for NeP in any cancer population among adults only.Studies were excluded if they assessed children or assessed neuropathy only for example, chemotherapy related peripheral neuropathy unless they specifically reported on painful neuropathy.
To identify articles that discussed the assessment of NCP, three search strategies were constructed for the databases Pubmed, Web of Science and Embase (Table 1).The search strategy was constructed around the following PICO (population-intervention-comparison-outcome) question: "What are the assessment methods (I) that are best suited to classify NeP (O) in the cancer population (P)?" Duplicated articles were removed with Endnote X8.2 and resultant articles consecutively imported into Rayyan where a screening on title and abstract was performed by N.M. and D.L.This screening assessed the agreement of the obtained articles with certain predefined inclusion and exclusion criteria (Table 2).Those articles that adhered to the inclusion criteria formed the basis for the recommendations on assessment of NCP.

| Search results
The three search strategies were entered into their respective database on the December 13, 2021, which resulted in 16,372 search results.After removal of the duplicates, 13,211 articles remained, of which 172 were eligible for inclusion based on title and abstract.One hundred and five articles were excluded primarily due for the following reasons: (1) did not examine components of NeP grading system (n = 63); and (2) were specific to chemotherapy-induced peripheral neuropathy (CIPN) without differentiating painful neuropathy (n = 42).The remaining 67 articles formed the base for the recommendations toward assessment that are formulated below.

| Assessment of possible NeP
Criterion 1: History of a relevant neurological lesion or disease and pain neuroanatomically plausible As NCP may arise from invasion of the cancer itself or as a consequence of treatment (chemotherapy, surgery and radiotherapy) the majority of cancer sufferers would fulfill the criterion of a possible lesion.Further guidance on establishing the possibility of a neurological lesion or disease includes consideration of self-reported symptoms (Finnerup et al., 2016).It is considered that symptoms such as burning, numbness, paraesthesia, electric shock or shooting pain are suggestive of, but not pathognomic for NeP (Finnerup et al., 2016).The DN4 and LANSS questionnaires have demonstrated good discriminative ability for NeP in both tumor-related and treatmentrelated NCP (Mulvey et al., 2017).
The following is a summary of symptoms consistently reported across people with NCP.

| Is pain neuroanatomically plausible?
When considering pain in a neuroanatomically plausible distribution guidelines advocate consideration of distribution within the innervation territory of a peripheral nerve or nerve root or the somatotopic representation of the body within the central nervous system (Finnerup et al., 2016).An additional consideration is added for pain distribution that may be different to this but is consistent with the disorder (Finnerup et al., 2016).This is particularly relevant for chemotherapy-induced NCP, which is consistently reported as distributed in a glove or stocking distribution, with milder cases associated with pain affecting a smaller or more peripheral area (e.g., fingertips) and more severe cases demonstrating symptoms extending further into upper and lower limbs (Attal et al., 2009;Beijers et al., 2014;Tofthagen et al., 2013;Wang et al., 2019).More generalized symptom distributions have also been described across cancer populations including within the head and neck and whole body have been reported (Attal et al., 2009;Tofthagen et al., 2013).
Comparing pain distribution among those with chemotherapy-related NCP versus other types of NCP within a mixed cancer population (n = 754), Coffeen et al. observed that glove/stocking distribution was more common in chemotherapy-related NCP (73%) versus other types of NCP (6%), while other types NCP had more variable patterns with 39% and 15% within tumor and surgery sites, respectively (Coffeen et al., 2019).In those considered to have surgical-related NCP, distributions recorded were more consistently within surgical sites and peripheral nerve distributions.In 506 people with breast cancer diagnosed with NeP using the NeP grading system (n = 506) 97% reported pain localized to surgical area, with only 3% in hands and feet (Pereira, et al. 2015).In a further 181 people (various cancer types) who underwent lymph node excision or sentinel node biopsy, symptoms were reported within distinct cutaneous distributions of peripheral nerves local to the surgical site (the medial, antebrachial cutaneous nerve, medial brachial cutaneous nerve and the intercostobrachial nerve for axillary surgeries; the femoral branch of the genitofemoral nerve, ilioinguinal nerve, lateral femoral cutaneous nerve, cutaneous branch of the obturator nerve and the anterior cutaneous branch of the femoral nerve for inguinal surgeries) (Høimyr et al., 2011).Finally, radiotherapy induced brachial plexopathy symptom distributions have been recorded within the ipsilateral arm and hand (Chen et al., 2012;Chen et al., 2014); within C5 and C6 distributions (Cai et al., 2016) and within the upper and lower trunk of brachial plexus (McNeish et al., 2020).
Recommendation: Clinicians and researchers should assess neuropathic symptoms and may screen using validated questionnaires such as the LANSS and DN4 which have demonstrated discriminative validity in cancer populations (Mulvey et al., 2017).Classic neuropathic symptoms present across all categories of treatment and tumor-related NCP; however, they appear to be more pronounced in chemotherapy-induced NCP.Further cold-evoked pain and cold-related symptoms are more evident in chemotherapy-induced NCP.
Assessment of distribution of symptoms relevant for NCP should include consideration of glove and stocking distribution (consistent with chemotherapy-induced NCP) as well as more classic cutaneous distributions associated with peripheral nerve, plexus or nerve root.

| Assessment of probable NeP
Criterion 2: Pain is associated with sensory signs in the same neuroanatomically plausible distribution Clinical examination is needed to support NeP classification.The presence of positive (i.e., signs of hyperalgesia/ allodynia) and negative (i.e., sensory loss) sensory signs are assessed and whether they follow a neuro-anatomically plausible distribution consistent with reported symptoms.
The presence of negative sensory signs is weighted more strongly than positive sensory signs (Boland et al., 2015;Finnerup et al., 2016), particularly if the latter do not follow a neuro-anatomically plausible distribution.Therefore, partial or complete loss to one or several sensory modalities (touch, vibration, cold or heat detection) can be regarded as more decisive and should be observed in regions that are neuro-anatomically plausible (Finnerup et al., 2016).The clinical examination can include a bedside examination, which can be complemented with quantitative sensory testing (QST).
A summary of bedside and QST assessments and nerve fibers assessed is presented in Table 3. Bedside examination should assess mechanical and thermal detection (touch, vibration, cold or heat detection) (Baron et al., 2010;Boland et al., 2015;Esin & Yalcin, 2014).QST is a psychophysical form of assessment for more accurate sensory profiling, conducted in specialist pain centers, which assesses sensory nerve function in a noninvasive way (Haanpää et al., 2011;Leone et al., 2016).
In those considered to have surgical-related NCP, negative sensory signs have been reported as the main sensory dysfunction after breast cancer surgery and associated with persistent postoperative pain (Andersen et al., 2017;Mustonen et al., 2020).This was demonstrated by increased thresholds for tactile detection threshold, mechanical pain threshold and thermal stimuli measured in the surgical area.A greater area of hypoesthesia was associated with an increased risk of pain and pain intensity.Vilholms et al. demonstrated that participants with breast cancer showed both negative (increased thermal thresholds) and more severe positive sensory signs (e.g., cold allodynia and temporal summation evoked by repetitive pinprick) on the operated side following surgery (Vilholm et al., 2009).Negative sensory signs have also been identified in people with multiple myeloma related peripheral neuropathy including distally distributed sensory hypoesthesia (mostly affecting feet) (Leone et al., 2016).In people with chemotherapyrelated NCP both negative and positive sensory signs are evident that is, increased thresholds to vibration, light touch and pinprick stimuli, as well as reduced heat and cold pain thresholds (Griffith et al., 2014;Kosturakis et al., 2014;Martland et al., 2020;Roldan et al., 2018;Zhi et al., 2019).Interestingly, multiple studies have found pre-existing subclinical sensory deficits prior to treatment in cancer patients, which are hypothesized to be disease driven, and which increased the risk for the later development of (severe) chemotherapy induced peripheral neuropathy (CIPN) (Attal et al., 2009;Boyette-Davis et al., 2012;de Carvalho Barbosa et al., 2014;Kosturakis et al., 2014;Reddy et al., 2016).The largest sensory deficits were observed distally, and less pronounced proximally.These observations indicate a correlation between apparent subclinical pretreatment neuropathy and an increased risk for developing CIPN, and suggest that chemotherapeutics may not be the only cause for neuropathy in patients receiving chemotherapy, but that the cancer disease process also contributes to neuropathy.
Recommendation: Clinical examination should combine several sensory modalities in testing to evaluate different nerve fiber types, for example, to assess touch and/or vibration (Aβ), pressure and or pinprick sensation (Aδ), cold (Aδ), and heat sensation (C fibers).These sensory responses should be compared between the symptomatic site and a nonpainful adjacent or contralateral site.Negative sensory signs (namely findings of partial or complete loss to one or several sensory modalities) can be regarded as decisive, while positive sensory gains are only considered supportive of NCP.Findings observed in regions that are neuro-anatomically plausible hold greater weight.Such changes in sensation can be assessed with a bedside examination.Applying QST might be required to obtain a more accurate sensory profile; however, this is typically conducted in specialist or research Centre.

| Assessment of definite NeP
Criterion 3: Diagnostic test confirming a lesion or disease of the somatosensory nervous system explaining the pain Several confirmatory tests have been proposed to help classify definite NeP (Table 4) and many have been utilized in the analyses of cancer populations, primarily in painful CIPN.

| Nerve conduction studies/ electromyography (EMG)
In CIPN, nerve conduction studies have been advocated as the gold standard assessment, which may be indicated in patients receiving platins and complaining of neuropathic symptoms posttreatment (Brozou et al., 2018).A progressive deterioration in the amplitude of sensory action potentials has been linked to chronic neuropathy (Argyriou et al., 2013).The sensory nerve action potential T A B L E 4 Diagnostic tests for lesion of the somatosensory nervous system

Nerve conduction studies
Nerve conduction studies are regarded as the gold standard for diagnosing large fiber neuropathies (Brozou, et al. 2018).Sensory conduction studies of sural and radial nerves are recommended for the diagnosis of mild, predominantly sensory axonal neuropathy (Rutkove, et al. 1997), and should be complemented with at least one motor study (typically of the tibial nerve), to assess for motor involvement (Brozou, et al. 2018).

Skin biopsy
The evaluation of intraepidermal nerve fiber density through skin biopsy is regarded as the most objective and sensitive test for small fiber neuropathy (Yoon & Oh, 2018) recommended by the European Federation of Neurological Societies (Lauria, et al. 2010).
Magnetic resonance imaging (MRI)/ computed tomography (CT) MRI and CT are commonly used to identify malignancy and can be used to assess neuropathy (Boland, et al. 2015).Diffusion weighted magnetic resonance neurography may offer more precise visualization of neural structures for example, brachial plexus (Andreou, et al. 2015).

Ultrasound
Ultrasound allows visualization of the nerve and is especially suited to assess the nerve volume (Briani, et al. 2013).

Laser-evoked potentials
Laser-evoked potentials provide a direct functional examination of the nociceptive afferents, and can provide a detailed sensory profile in neuropathic pain (Valeriani, et al. 2012) changes with progressive neuropathy, enabling nerve conduction studies to detect differences in CIPN severity in an objective manner (Matsuoka et al., 2016).Further, reduced dorsal sural nerve sensory action potential amplitude has been shown to be the first, albeit mild and sub-clinical, neurophysiologic alteration in myeloma patients with painful CIPN (Dalla Torre et al., 2016).Sensory axonal excitability abnormalities have also been shown to be predictive of final clinical outcomes on an individual patient basis in painful CIPN (Park et al., 2009).Therefore, sensory axonal excitability may enable identification of preclinical CIPN prior to the onset of chronic neuropathy and may provide a sensitive biomarker for the severity of neurotoxicity.Additionally, abnormal spontaneous high frequency motor fiber activity has been shown to be a sensitive and specific endpoint of acute oxaliplatin-induced motor nerve hyperexcitability, detectable on EMG on days 2-4 posttreatment (Hill et al., 2010).In people with multiple myeloma, nerve conduction studies have revealed bilateral and symmetric amplitude reduction of sural-nerve sensory action potentials; whereas the other variables (e.g., compound motor action potentials and conduction velocities) were spared (Leone et al., 2016).These findings were obtained in both patients with painful (82%) as non-painful (18%) multiple myeloma.Nerve conduction/electromyography is also useful for investigation of NCP related to other aetiologies.For example, in people with metastatic brachial plexopathy two studies report both upper and lower brachial plexus trunk involvement evident on EMG (Kim et al., 2020, McNeish et al., 2020) While safe and noninvasive, nerve conduction studies can be uncomfortable and as such their use in a population with high investigation/treatment load should be carefully considered.Further, while research indicates their use in monitoring progression of (painful) CIPN, self-report measures and clinical evaluation are much more commonly used in practice.

| Skin biopsy
The evaluation of intraepidermal nerve fibers through skin biopsy is regarded as the most objective and sensitive test for small fiber neuropathy and has received a level A recommendation from the European Federation of Neurological Societies to measure the density of small fiber epidermal innervation (Lauria et al., 2010).Broadly, a decrease in intraepidermal nerve fiber density has been shown to correlate with NeP symptoms and the results of other neurophysiological tests (Chan & Wilder-Smith, 2016).Results from skin biopsies in people with CIPN seem to vary.For example, some studies show it was not a clinically relevant objective measure of peripheral nerve deterioration during oxaliplatin treatment (Bechakra et al., 2018;Krøigård et al., 2020;Velasco et al., 2017), nor specifically related to symptoms (Koskinen et al., 2011).However, subepidermal nerve fiber density and corneal nerve fiber density have been shown to be significantly decreased, while axonal swelling ratio and upper dermis nerve fiber density have been shown to be significantly increased in people with painful CIPN (Bechakra et al., 2018;Ferdousi et al., 2015).Interestingly serum nerve growth factor has been shown to be increased after treatment in people displaying painful CIPN, compared with those with painless or absent CIPN and was significantly correlated with NeP (Velasco et al., 2017).

| MRI and CT
Several imaging techniques, including CT and MRI, are currently recommended for the assessment of cancerrelated neuropathy and NCP (Boland et al., 2015).The use of CT and MRI are valuable in the identification of tumor growth compressing a neural structure and such observations are regarded as objective confirmatory tests.Additionally, they provide evidence for a history of a relevant disease/lesion.In people with metastatic brachial plexopathy, two studies reported both upper and lower brachial plexus trunk involvement (Kim et al., 2020, McNeish et al., 2020), with one demonstrating greater upper trunk involvement than lower trunk involvement (Kim et al., 2020).Recent advances to standard MRI of the peripheral nerves and plexi include use of diffusion weighted magnetic resonance neurography (Andreou et al., 2015;Chalasani et al., 2020).For example, 12 out of 40 participants with a history of malignancy and symptoms of brachial plexopathy demonstrated evidence of brachial plexopathy of different aetiologies (10 malignant plexopathy; 1 postradiation neuritis, 1 neuroma) using this method.Further, MR neurography has benefits of being able to identify dorsal root ganglia hypertrophy, which has been found to be a significant morphological correlate of sensory neuropathy (Apostolidis et al., 2017).

| Ultrasound
Ultrasound offers the ability to visualize nerves and assess cross-sectional area/nerve volumes in a noninvasive way.Mixed findings have been reported in cancer populations.One study did not reveal decreased cross-sectional area, commonly found in axonal neuropathies, but instead reported an increased cross-sectional area at entrapment sites (median nerve at wrist and ulnar nerve at elbow) in 60% of people with CIPN (Briani et al., 2013).Another study reported the sural nerve cross-sectional area to be smaller in people with CIPN versus healthy controls (Lycan et al., 2020).In addition, time since chemotherapy was associated with smaller sural nerve CSA, and a decrease in sural nerve cross-sectional area was associated with a decrease in the distal intraepidermal nerve fiber density (Lycan et al., 2020).As such, ultrasound may be useful but further research is warranted before recommendations can be made.

| Laser-evoked potentials (LEP)
LEP recordings offer the main advantage of providing a direct functional examination of nociceptive afferents fibers, and can provide a detailed sensory profile in NeP (Valeriani et al., 2012).Previous studies in polyneuropathy have shown that the presence of ongoing pain is related with LEP suppression, a result of extensive axonal damage (of Aδ-fibers), while hyperalgesia or allodynia correlates with partially preserved LEPs, suggestive of Aδ partial denervation plus sensitisation (Truini et al., 2010).In CIPN suppressed LEPS have been identified in both upper and lower extremities (Isak, et al. 2021).Furthermore, evidence of abnormalities in cutaneous silent periods (more than in LEPs) in CIPN may indicate greater peripheral nociceptive fiber pathology versus combined peripheral and central nociceptive processing.LEPs carry limitations in that they can only be assessed in dedicated pain clinics or research laboratories, can cause skin irritation, and are not suited for the functional investigation of C-fibers (Valeriani et al., 2012;Yoon & Oh, 2018).
Recommendation: To classify definite NeP, at least one confirmatory test should be applied.MRI/CT are common investigations for both malignancy and identification of lesions of the somatosensory nervous system.Nerve conduction studies have the strongest evidencebase, followed by skin biopsy for specific classification of neuropathy.However, as these procedures can cause discomfort, are each limited to particular nerve fiber types or can only be conducted in specialist centers, clinicians should carefully consider the balance of benefit to the patient, particularly in a cancer population with such a high investigation and treatment load.

| CONCLUSION
This narrative review provides a summary of the research evidence for the presence of subjective, psychophysical and objective features according to the updated NeP grading system among populations with cancer-related pain.While self-report measures for both NeP and neuropathy are commonly used in cancer populations, incorporating assessment that aligns with the NeP grading system is important.This might commence with the use of NeP screening questionnaires, then proceed to evaluate the nature and distribution of pain and related symptoms to establish possible NeP.Clinical sensory testing for negative and positive sensory signs is required for a classification of probable NeP and may be complemented by QST where available.Confirmation of a lesion of the somatosensory system may then be achieved from a range of diagnostic tests.
A number of important considerations emerge from the evidence for assessment of NCP.The first is consideration of distribution of pain and sensory signs.In chemotherapy-related NeP, this would involve glove and stocking distribution extending into the limbs with more severe neuropathy.Further, cold-evoked symptoms appear to be more common in this sub-group.Ideally diagnostic testing is undertaken for confirmation of NeP; however, careful considerations of the benefits to the patient are warranted given the assessment and treatment load already experienced by many individuals with a cancer diagnosis.Finally, once possible, probable or definite NCP has been established, onward referral for appropriate pain management (medical and allied health) services is warranted.