NCP is pain caused by direct damage to the nervous system. A nerve can be infiltrated or compressed by a tumor, or strangulated by fibrosis [4]. Once a peripheral nerve is damaged, the pain fibers become abnormally sensitive, triggering spontaneous pain that is amplified in the spinal cord; even a minor stimulus such as a touch can trigger pain (allodynia). The pain may persist for months or years after damaged tissue has healed. In such a setting, pain no longer reflects ongoing injury but, rather, a malfunctioning nervous system. NCP is part of a larger spectrum of neuropathy caused by the cancer per se, treatment, paraneoplastic syndrome, and comorbidities such as diabetic polyneuropathy and postherpetic neuralgia. Here, we focus on NCP caused by cancer per se and chemotherapy-induced peripheral neuropathy (CIPN) [5]. NCP is one of the three main types of cancer pain (the others are somatic and visceral pain). There are two principal forms of pain pathophysiology: nociceptive and neuropathic. NCP is nerve-related (typically neuron-related) pain characterized as a burning or electrical sensation; however, NCP sometimes manifests as decreased sensation or actual muscle weakness [6]. In clinical settings, both hypersensitivity and hyposensitivity symptoms often coexist. NCP is often a part of a mixed syndrome with other types of pain including somatic or visceral pain. NCP is chronic and often manifests as persistent background pain with acute exacerbations of breakthrough pain several times daily. The breakthrough pain is often spontaneous, but can also be triggered by movement, touch, cold, and heat. Such spontaneous or triggered pain is perceived as a sensory abnormality (i.e., hyposensitivity, hypersensitivity, or both), although paresthesia (an abnormal sensation, such as tingling, an electric shock, or burning) may also be considered a typical hypersensitivity symptom. Allodynia is a type of pain evoked by non-painful stimulation, such as a touch; hyperalgesia (increased response to a stimulus that is not normally painful) is often evident. Hyposensitivity symptoms such as diminished or absent cutaneous sensations often present with balance/gait disturbances. Reduced sensation compromises fine motor skills such as the use of chopsticks or buttoning of shirts.

In contrast, nociceptive pain is the result of injury to somatic and visceral structures, followed by activation of nociceptors in skin, the viscera, muscles, and connective tissue. Nociceptive pain can be divided into somatic and visceral pain. Somatic nociceptive pain is sharp, well-localized, throbbing, and pressure-like. However, visceral nociceptive pain developing secondary to compression, infiltration, or distention of the abdominal or thoracic viscera is often described as more diffuse, less localized, ache-like, and cramping in nature [4]. Table 1 summarizes the differences between neuropathic and nociceptive pain in terms of pathophysiology, symptoms and signs, duration, and impact on the quality-of-life.

Some types of NCP, such as spinal cord compression accompanied by motor and sensory changes, and tumor infiltration of the cord confirmed by imaging, are definitive. However, other NCPs are mixed with somatic or visceral nociceptive cancer-related pain; appropriate diagnostic standards remain to be established. Thus, few prospective data on the prevalence of NCP and the underlying mechanism(s) are available. An extensive meta-analysis of over 10,000 patients was published in 2012; the prevalence of NCP, both pure and mixed, was approximately 39% [3]. The prevalence rates of pure neuropathic pain, mixed pain, and pure nociceptive pain are approximately 19%, 20%, and 59%, respectively. The prevalence of NCP has been reported to be as high as 40% among cancer patients. The cited meta-analysis explored NCP etiologies. Sixty-four percent of NCP was caused by cancer per se, and 20% by treatments such as chemotherapy, radiotherapy, and cancer surgery. Notably, these proportions differed from those of all cancer pain caused by treatment (only 10%), suggesting that a higher proportion of NCP was caused by treatment(s).

A cross-sectional Dutch study reported that 23% of oncology outpatients (204/892) experienced moderate-to-severe pain [7] and 19% exhibited neuropathic symptoms (170/892). However, the NCP prevalence increased to over 40% in patients with moderate-to-severe pain, affecting the activities of daily life more severely than was the case for patients without neuropathic pain (mean interference score 4.7 vs. 4.17, p = 0.054). Even patients with mild neuropathic pain experienced significant interference with the activities of daily life. Despite the high prevalence and severity of such pain, only 8% were taking adjuvant analgesics.

Recently, a Korean, multicenter, cross-sectional observational study explored the prevalence and impact of NCP on the quality-of-life of cancer patients [8]. The prevalence of NCP of score ≥ 1 on a visual analog scale (VAS) was 36.0% (722/2,003). Approximately 20% of all patients reported moderate-to-severe pain. Importantly, those with VAS scores ≥ 4 exhibited a higher prevalence of NCP than did those with mild pain (42.4% vs. 27.4%). Cancer patients with NCP reported more high-level pain, more severe interference with the activities of daily life, and a poorer quality-of-life, than did those without NCP. However, only 49.6% of patients with moderate-to-severe NCP received adjuvant analgesics including antidepressants and anticonvulsants. Similarly, the NCP prevalence in a Western study was approximately 40%; both the quality-of-life and the extent of interference with daily activities were more severely affected in patients with than without NCP [9]. In both Korea and Western countries, adjuvant drugs are under-prescribed.

In summary, the prevalence of NCP remains high, interfering significantly with the activities of daily life, and current management strategies are inadequate [3,7-10].

NCP can be caused by direct cancer invasion or nerve compression, and/or cancer treatments including surgery, chemotherapy, and radiotherapy. NCP is divided into several categories: plexopathy, radiculopathy, peripheral neuropathy, paraneoplastic sensory neuropathy, leptomeningeal metastasis, cranial neuralgia, and malignant painful radiculopathy (Table 2) [5,11,12]. Awareness of the various types of NCP improves diagnosis, treatment, and outcomes. Cancer pain is often of a mixed nature or, if it is purely neuropathic, may be one of several pains experienced. NCP caused by the tumor per se usually involves both nociceptive and neuropathic components, and mixed pain is more common than NCP caused by cancer treatments [13]. Most NCP caused by chemotherapy is purely neuropathic in nature [14]. Various surgeries, such as mastectomy, neck lymph node dissection, laparotomy and thoracotomy, cause neuropathic pain. Occasionally, surgery-related NCP can become chronic, although improving over time.

CIPN is a common problem in cancer patients because survival has been significantly prolonged by advances in chemotherapy; however, neither the level of awareness of CIPN nor the management thereof is adequate. CIPN affects both the quality-of-life and eventual outcomes because of treatment delays, dose reductions, and drug discontinuation [15]. Many chemotherapeutic agents can cause CIPN, of which the most common are platinum agents (cisplatin, oxaliplatin, and carboplatin); the taxanes (taxol and docetaxel); and the vinca alkaloids (vincristine and vinorelbine) [16,17]. The symptoms range from early post-treatment pain, such as paclitaxel induced acute pain syndrome (P-APS) [18-20] and cold allodynia after oxaliplatin infusion, to chronic peripheral sensory neuropathy featuring tingling, pain, and decreased sensation [21-23]. CIPN is usually a dose-dependent, cumulative side-effect exhibiting a glove-stocking distribution. CIPN symptoms include sensory loss, paresthesia, dysesthesia and pain, sometimes accompanied by muscle weakness [16,24-26].

Paclitaxel is a chemotherapeutic agent used to treat breast, ovarian, lung, and head-and-neck cancer, and is associated with two types of neuropathic pain: P-APS (acute type) and a classical peripheral neuropathy (chronic type). P-APS occurs within days after each dose and usually also abates within days; it was earlier termed paclitaxel-induced arthralgia/myalgia, and has recently been named P-APS. This syndrome, described in up to 70% of patients, usually develops within 1 to 3 days of paclitaxel administration; however, the symptoms largely resolve within 7 days [19,24,26].

Oxaliplatin is a new-generation platinum compound widely used to treat various cancers including colorectal, stomach, and pancreatic cancers. Oxaliplatin-induced peripheral neuropathy is an acute prickly dysesthesia affecting the hands and feet, combined with spontaneous pharyngolaryngeal dysesthesia often triggered by cold. The conditions develop during oxaliplatin infusion and subside within few days. Oxaliplatin-induced early acute dysesthesia occurs in up to 90% of patients. Simply opening the door of a refrigerator or drinking cold water can induce a painful prickly sensation. Such oxaliplatin-induced, acute cold allodynia resembles P-APS. On repeated infusion, oxaliplatin-induced CIPN can include decreased feeling in the hands and feet, impairing balance and the sensing of vibration and touch [21,22].

CIPN mechanisms include disruption of axonal transport, changes in ion channel and receptor activities, neuronal injury and inflammation, oxidative stress, and mitochondrial damage [16]. Taxane, platinum, and vincristine affect the peripheral sensory neurons of the dorsal root ganglia (DRG) and the spinal cord; they are toxic to DRG neurons, and increase the activities of both voltage- and ligand-gated ion channels (the sodium, calcium, and transient receptor potential channels). Despite the common pathogenic pathways, taxane and vincristine trigger inflammation of the dorsal horn of the spinal cord more strongly than do platinum compounds. Taxane and vincristine activate cord microglia, astrocytes, and satellite glial cells, triggering release of tumor necrosis factor and interleukin-1β.

Unfortunately, CIPN can persist for several months to years, even after discontinuation of chemotherapy, and may never be completely eliminated. The symptoms and durations of CIPN are summarized in Table 3.

Pain is very subjective and it is not always easy to approach patients complaining of pain. If NCP is diagnosed, the pain should have the characteristics of neuropathic pain, must be caused by a distinct lesion or disease of the somatosensory system, and should be assessed using electrophysiological tests to determine objectively the extent of nerve damage.

Conventional nerve conduction and evoked potential studies assess only large myelinated nerve fibers; they do not detect damage to small fibers. Laser-evoked potentials stimulate skin Aδ fibers and can be used to assess the integrity of the peripheral free nerve endings of the sensory cortex [37]. However, the technique is available only in dedicated pain clinics or research laboratories. QST can be employed to assess the functionality of both large and small fibers. QST evaluates both negative symptoms and positive sensory phenomena by measuring the thresholds of pressure, the pinprick test, and vibration, as well as cold and warm perception [38]. However, the QST is a psychophysical test that often generates false-positive results, is time-consuming, and cannot localize lesions causing neuropathic pain.

Evaluation of intraepidermal nerve fibers (IENFs) via skin biopsy is the most objective and sensitive test for neuropathic pain. The diagnostic accuracy of the QST used to assess small fiber neuropathy is low (approximately 50%), but IENF quantification is highly sensitive and specific (up to 90%) [39]. Skin samples obtained via 3-mm punch biopsy are stained for protein 9.5, a pan-axonal marker. The number of stained Aδ fibers/mm that cross the dermal-epidermal junction is the IENF density, a decrease in which correlates with neuropathic symptoms and the results of other neurophysiological tests [40]. However, the test is available only in specialized centers.

The tricyclic antidepressants (TCAs) have been used for decades. TCAs (amitriptyline, nortriptyline, and desipramine) demonstrated efficacy for moderate non-cancer neuropathic pain such as diabetic polyneuropathy, postherpetic neuralgia, migraine, and fibromyalgia [42]. Amitriptyline or nortriptyline are old but potent drugs. Although a recent Cochrane review found that TCAs afforded at least moderate pain relief (number needed to treat = 3.6) in patients with nonmalignant neuropathic pain [43], studies on cancer patients revealed only minimal analgesic effects. The cardiac side-effects (QT prolongation, orthostatic hypotension) and anticholinergic effects (confusion, sedation, dry mouth, somnolence, and bladder distension) limit the use of these drugs in older patients.

Newer antidepressants, the SNRIs, have fewer side-effects than TCAs. Unfortunately, serotonin-selective reuptake inhibitors (SSRIs) do not relieve pain. In a randomized, double-blind crossover trial on 231 patients with taxane- or platinum-treated, cervical intraepithelial neoplasia, duloxetine (30 mg daily for 1 week followed by 60 mg daily for 4 weeks) decreased the average pain score by 1.06, compared to 0.34 in the placebo group [44]. Fifty-nine percent of patients receiving duloxetine reported some pain reduction. A recent study showed that the drug relieved musculoskeletal pain in patients with early-stage breast cancer [45]. The adverse effects are those of all SNRIs, including nausea, dizziness, headache, and dry mouth. The drug is not recommended for patients with hepatic disease or severe renal impairment (CrCl < 30 mL/min). Venlafaxine (another SNRI) relieves diabetic neuropathic pain and fibromyalgia. A randomized, double-blind, placebo-controlled phase 3 study reported that venlafaxine was clinically active against oxaliplatin-induced, acute neurosensory toxicity [46]. However, efficacy in patients with chronic CIPN has not yet been proven. The National Comprehensive Cancer Network (NCCN) guidelines recommend venlafaxine and duloxetine, the newest SNRI [47], as the first choice for NCP patients; both drugs are effective against diabetic neuropathic pain. Also, SNRIs relieve the depression and anxiety accompanying chronic cancer pain. However, several weeks are required before adjuvant analgesic effects are evident, and somnolence, dry mouth, dizziness, and increased sweating are common side-effects.

Anticonvulsants have been used since the 1970s to manage pain. The World Health Organization and NCCN guidelines suggest that adjuvant analgesics should be combined with opioids to relieve cancer pain. Pregabalin and gabapentin have been proposed as first-line treatments for cancer-related neuropathic pain; these new-generation anticonvulsants have fewer side-effects than older drugs such as carbamazepine and valproic acid. Anticonvulsants act by blocking the sodium channels of the affected peripheral nerves, reducing spinal, glutamatergic nociceptive transmission and enhancing the descending inhibition of spinal nociceptive transmission.

Gabapentin is currently widely used because several clinical trials demonstrated efficacy against NCP when the drug was given either alone or in combination with opioids. Pregabalin has the same mechanism of action and the same indications. This drug binds to the α2δ subunit of the voltage-gated calcium channel, reducing the levels of excitatory neurotransmitters. Although the drugs were effective against neuropathic pain of various etiologies, they have not been shown to be effective against NCP. Only one randomized controlled trial found that pregabalin was more effective, and required a lower level of a combined opioid, than gabapentin and amitriptyline, associated with a significantly lower mean VAS score [48]. However, pregabalin did not reduce the chronic pain associated with oxaliplatin therapy [49]. Gabapentinoids usually reduce anxiety and sleep disturbances in those with chronic pain and can also be helpful in patients whose pain does not respond to opioids or for whom opioid reductions are required. The maximal daily doses are 3,600 mg/day for gabapentin and 600 mg/day for pregabalin; the latter drug exhibits linear pharmacokinetics and dose-dependent absorption, thus reducing pain rapidly and is six-fold more potent than gabapentin. Common side effects include somnolence, dizziness, and edema. Renal dose adjustment is required for both drugs.

Carbamazepine, a sodium channel blocker, is used to treat trigeminal neuralgia. Common side-effects include dizziness, skin rash, hyponatremia, and leukopenia. The drug has not been proven to act against NCP, but can be cautiously used by those with neuralgic pain caused, for example, by cancer infiltration of peripheral nerves.

Neuropathic pain responds very poorly to conventional analgesics such as opioids, paracetamol, and nonsteroidal anti-inflammatory drugs. Both the NCCN and World Health Organization guidelines recommend that combinations of opioids and adjuvant analgesics should be used to treat mixed or somatic pain even in patients with mild-to-moderate cancer pain. A patient may simultaneously experience mixed visceral and neuropathic pain in one lesion and pure somatic or neuropathic pain elsewhere. Therefore, opioids are used to control cancer-related neuropathic pain. However, NCP is not effectively controlled by opioids only. Adjuvant analgesics such as antidepressants and/or anticonvulsants are required [50]. Although combinations of opioids and adjuvant analgesics help to control NCP, the situation remains problematic, both in terms of pain relief per se and the management of side-effects. Moreover, the prescription rate of adjuvant analgesics to treat NCP is no higher than 20% worldwide.

Topical 5% (v/v) lidocaine and capsaicin (8% w/v) patches effectively treat postherpetic neuralgia [51]; however, their effects on NCP remain unclear. Skin patches may be useful in patients who cannot take oral medication or who exhibit dynamic allodynia as the patches prevent skin stimulation. Based on the Grading of Recommendations Assessment, Development and Evaluation (GRADE) classification, drugs that may not be useful for patients with neuropathic pain include lacosamide, lamotrigine, oxcarbazepine, topiramate, and zonisamide [50]. Levetiracetam and mexiletine are not recommended. Apart from pharmacological management, spinal cord stimulation, scrambler therapy or even acupuncture may afford relief, and are worth exploring.