Thus far, nerve enlargement identified by NMUS is mainly reported in the context of demyelinating polyneuropathies, such as immune-mediated neuropathies, while in axonal polyneuropathies focal or generalized nerve enlargement is an uncommon finding [25]. CIPN is characterized as (predominantly) axonal neuropathy mostly affecting sensory nerves [1].

In addition, we performed a systematic search in PubMed for relevant articles on NMUS and chemotherapy- or immunotherapy-induced polyneuropathy in humans. Articles that were published until March 2023 and written in English were encluded. We used the following search terms: “neuropathy”, “CIPN”,” ‘chemotherapy-induced peripheral neurotoxicity’, ‘toxic neuropathy’, ‘peripheral neuropathy’ and ‘neurotoxicity syndromes’ combined with “cytostatic”, “chemotherapy”, “immunotherapy”, “immune therapy”,”immune checkpoint inhibitor”, “immune checkpoint inhibition”, “checkpoint inhibition”, “checkpoint inhibitor” and “ultrasound”, “sonography”, “nerve sonography”, “nerve ultrasound”, “neuromuscular sonography”, “NMUS” and “HRUS”. There were no exclusions due to article type. We identified 6 relevant publications (1 case report, 1 review, 1 retrospective study, 3 prospective studies) reporting on cancer patients who were treated with cytostatic agents. No publications regarding NMUS use for neuropathies induced by immunotherapies were identified. On account of the limited number of publications and patients included, a meta-analysis was not feasible. The basic characteristics of the analysis are summarized in Table 2.

As shown in our systematic literature search we identified a case report as well as a few monocentric retrospective and prospective studies showing CSA increase of multiple nerves in patients with chemotherapy-induced polyneuropathy. This is consistent with our clinical experience and the case studies presented in this work.

In patient 1, atypical presentation of cranial nerve involvement and a severe affection of motoric functions led to extensive diagnostics to exclude leptomeningeal disease, infectious and autoimmune CNS/PNS affection of other causality. After excluding concurrent conditions and based on the additional NMUS findings, an interdisciplinary consent led to discontinuation of epirubilin. The increased CSA of the cervical dorsal roots and the pattern of continuously enlarged multiple nerves did not fit any known constellation, e.g., CIDP. This supported the hypothesis of therapy-induced neuropathy, potentially with an accompanying inflammatory component as described in some previously published cases of CIPN. Neither electroneurography nor CSF and extensive blood testing were diacritic to confirm neuropathy induced by cancer treatment. Patient 1 exhibited a significant reduction of the neurological symptom burden after only a few weeks of treatment interruption. At the same time, NMUS pathologies improved stepwise (without any concurrent therapeutic intervention besides physiotherapy/occupational therapy). In this case, NMUS has proven to be a very helpful tool for both the diagnosis and neuromonitoring of neurotoxicity compared with neurography findings, which showed improvement in a delayed fashion. This case likewise illustrates the chance of symptom improvement even in cases of severe neuropathies within an interdisciplinary expert setting.

In patient 2, the additional NMUS assessment showed a strong PNS inflammation where a severe CNS inflammation syndrome was dominating the clinical decision making. As the patient was unable to cooperate in a sufficient manner or describe the subjective sensorimotor symptoms initially, NMUS contributed to differentiate immune-related inflammation of the nervous system from other conditions. Therefore, the technique seems to be helpful in distinguishing between peripheral and central causes of muscular weakness in cases of impaired communication. According to this, the value of NMUS as a specific biomarker of peripheral nirAEs should be investigated in future prospective studies.

In our patients, neuromuscular ultrasound revealed increased CSA in several nerves with partly continuous and partly discontinuous swelling, which is consistent with a major part of the reviewed studies. Moreover, in both patients the CSA of the dorsal roots and the vagus nerve in patient 2 were enlarged, which has not been reported in treatment-induced neuropathy so far. Both patient cases show the importance of extensive clinical and paraclinical neurological assessment in case of newly occurring neurological symptoms in systemically treated cancer patients. Furthermore, we emphasize the role of specialized neurologists or neuro-oncologists in this extremely interdisciplinary field for decision making in cases with such atypical clinical constellations of suspected neurotoxicity.

All previous studies, however, recruited patients with subjective suffering from neuropathy symptoms. Thus, it is currently not possible to discuss NMUS findings in cancer patients receiving neurotoxic agents but without relevant neurological symptoms. Most identified studies reported CSA increase of median nerves. Further nerve swelling was described in ulnar, radial, tibial and fibular nerves mostly presenting as concurrent enlargement of multiple nerve sites. Portland et al. performed a NMUS of both median nerves only, so a statement regarding the general swelling pattern in this patient case is not possible [26]. A potentially increased sensitivity to treatment-induced neuropathies in cancer patients with pre-existing compression syndromes such as carpal-tunnel-syndrome should be considered. All identified studies contained small groups of patients (< 30 patients each). Therefore, only descriptive statistical analysis was possible. The study designs regarding measured parameters and control groups were heterogeneous impeding a comparison and generalization of the presented data. Historically, healthy controls and CSA analysis were used in the prospective study of Lycan et al., whereas Erdmann et al. compared CIPN with CIDP and critical illness neuropathy patients without a healthy control cohort and regarding nerve echogenicity only without CSA assessment data presented in the manuscript [27, 28]. Moreover, we had difficulties retrieving information about the quantitative NMUS scores being used for the single cohorts in the studies.

A standardized neurological assessment of cancer patients with solid or haematological diseases who develop new neurological symptoms during or shortly after the treatment with neurotoxic agents is essential for early recognition of treatment-associated effects. To date, there is no effective preventive procedure for CIPN or immune-related neuropathies although interventions regarding dietary recommendations, medication and exercise are currently under investigation [32]. Early identification and neuromonitoring may lead to sufficient symptom control, more personalized cancer treatment and dose adjustment while avoiding long-term quality of life impairment [33, 34]. Moreover, the knowledge of clinical patterns of neurotoxicity is the key for the challenging differentiation between treatment-induced neuropathies, paraproteinemic conditions, leptomeningeal spread mimicking a polyradiculitis as well as “classical” neurological diseases (amyotrophic lateral sclerosis (ALS), multiple sclerosis MS, GBS/CIDP) in cases with atypical clinical presentation. In addition, there is an urgent need for valid biomarkers and pain-free, cost-effective techniques to differentiate treatment-induced neuropathy from other types of neuropathies and to monitor the severity as well as treatment response in terms of modern combinations of cytostatic and immunotherapeutic agents.

In summary, this is the first description of enlargements of dorsal roots and vagus nerve in cancer patients with treatment-induced neuropathies. In addition, we present a diagnostic algorithm that incorporates NMUS (as shown in Fig. 2) as an integral aspect of the neurological assessment for cancer patients. The algorithm may be particularily valuable as a part of the routine neurological assessment for cancer patients with suspected peripheral neurotoxicity in case of atypical presentations, severe symptoms, or acute onset, aiding in the process of differential diagnosis.

Diagnostic algorithm for cancer patients presenting with suspected treatment-related peripheral neurotoxicity. The figure shows our propose for a standardized diagnostic algorithm including NMUS as a part of routine neurological assessment for cancer patients with suspected peripheral neurotoxicity. Abbreviations: CNS central nervous system, CSF cerebrospinal fluid, MRI magnet resonance imaging, IgG immune globulin G, Auto AB auto antibodies, anti MAG antibodies against myelin-associated glycoprotein, PNP polyneuropathy, PNS peripheral nervous system

However, this current work shows representative examples from interdisciplinary NMUS integrative work-up only and does not allow any generalization of the identified NMUS findings. Taken together with previous studies, our experience nevertheless supports the promising potential of NMUS in this field and the importance of prospectively designed, cross-sectional studies for cancer patients with newly occurring neuropathies.