Role of magnetic resonance neurography in intercostal neuralgia; diagnostic utility and efficacy

Abstract

Objective:

To evaluate the utility and efficacy of MR neurography (MRN) in the diagnostic work-up for intercostal neuralgia and to assess the treatment course and outcomes in MRN-imaged clinically suspected intercostal neuropathy cases of chronic chest and abdominal wall pain syndromes.

Methods:

Following a retrospective cross-sectional study, a consecutive series of patients who underwent MRN of torso for suspected intercostal neuralgia were included. Patient demographics, pain location/level/duration, previous work-up for the same indication, MRN imaging results, and MRN cost per patient were recorded. An inter-reader reliability assessment was performed on the MRN findings using Cohen’s weighted κ analysis. Post-MRN treatment choice, as well as success rates of MRN directed perineural injections and surgical management were also evaluated.

Results:

A total of 28 patients (mean ± SD age, 48.3 ± 18.0 years, female/male = 3.0) were included. Pain and/or numbness in the right upper quadrant were the most common complaints. The mean maximum pain level experienced was 7.4 ± 2.5 on a 1 (lowest pain level) - 10 (highest pain level) visual analog scale. The duration of pain before MRN work-up was 36.9 ± 37.9 months. The patients had seen an average of 5 ± 2.8 physicians for such syndromes. 20 (71%) patients had one or multiple other imaging studies for prior work-up. MRN identified positive intercostal nerve abnormality in 19 cases with clinical symptoms of intercostal neuralgia. From the inter-reader reliability assessment, a Cohen’s weighted κ value of 0.78 was obtained. The costs of work-up was about one-third with MRN for diagnostic purposes with less financial and psychological harm. Among the MRN-positive cases, 9/19 patients received perineural injections, of which 6 reported improvement after their first round, lasting an average of 41.1 ± 83 days. Among the nine MRN-negative cases, two received perineural injections, of which none reported improvement. Surgical management was mostly successful with a positive outcome in six out of seven operated cases (85.7%).

Conclusion:

MRN is useful in diagnostic algorithm of intercostal neuralgia and MRN-positive cases demonstrate favorable treatment response to perineural injections and subsequent surgical management.

Advances in knowledge:

The use of MRN in intercostal neuralgia is an application that has not been previously explored in the literature. This study demonstrates that MRN offers superior visualization of pathology in intercostal neuralgia and confirms that treatment directed at MRN identified neuropathy results in good outcomes while maintaining cost efficiency.

Introduction

Intercostal neuralgia is a sensory neurogenic condition defined as an intense, sharp, shooting or burning pain, with or without hyperesthesia / paresthesia in the distribution of one or more intercostal nerves. ¹ In clinical practice, it usually falls under the umbrella of two common pain diagnoses – chest wall pain syndrome (notalgia paresthetica) and chronic abdominal wall pain syndrome. Chest wall pain symptoms account for up to 50% of all complaints in the ambulatory and emergency room settings, and chronic abdominal wall pain symptoms have been reported to account for up to 30% of chronic abdominal pain syndrome cases with mostly negative findings on prior diagnostic work-up. ^2,3 While there are many causes of chest wall pain and chronic abdominal wall pain syndromes, intercostal neuralgia is an important cause that remains relatively underdiagnosed and undertreated due to the overlap of symptoms with more common visceral causes of chest and abdominal pain, inadequate imaging modalities for identification, and decreased awareness of the condition. ^4–6 As such, patients will often undergo extensive and even invasive diagnostic work-up including abdominal CT, upper gastrointestinal endoscopy/colonoscopy, nuclear medicine functional biliary scans, echocardiography, cardiac angiography, laparoscopies and laparotomies to rule out underlying possible visceral, cardiac, or pulmonary sources of pain. ^5,7 In the event, if no direct cause of pain is found, the patients will often receive a final diagnosis of functional pain or conversion disorder. ⁵ Such course of care for this patient population results in extensive health-care costs with little to no help to the patient’s sufferings. ⁸ The diagnosis of intercostal neuralgia is heavily dependent on obtaining detailed patient history and performing a focused physical exam, which could still be subjective and non-specific. ⁴ Thus, it is essential to develop an appropriate cost-effective diagnostic approach for intercostal neuralgia as well as a treatment plan that can effectively decrease pain and neurogenic symptoms for such patients and improve their quality of life and daily functioning.

For management purposes, an intercostal nerve block using local anesthetics with or without a corticosteroid is usually performed to relieve the pain. ⁹ The accuracy and reliability of the nerve block can be further increased using ultrasound, fluoroscopy or CT guidance. ^10,11 If the pain is relieved partially or completely and sustained for some time post-injection, the diagnosis can be confirmed. Repeat therapeutic injections might be necessary to manage the patient’s pain. However, considerable placebo effect might be encountered with these injections and, additional or underlying psychological component contributing to the pain syndrome is also challenging to exclude clinically. Other treatment modalities include physical therapy, surgery (neurolysis/neurectomy), epidural injections, and selective radiofrequency nerve ablation/cryoablation. ^12,13

In recent years, magnetic resonance neurography (MRN), i.e. dedicated MRI of peripheral nerves has gained popularity among clinicians due to its ability to clearly depict peripheral nerve anatomy and pathology, therefore allowing clinicians to more accurately identify the etiology of patient’s pain among other similarly presenting differentials. MRN technology has been shown to aid the diagnostic work-up of a variety of neuralgias and pain syndromes such as cauda equina syndrome, lumbosacral plexus neuropathies, groin and genital neuralgias, and pelvic pain. ¹⁴ Due to its effective vascular signal suppression, MRN can be used to visualize the fine abdominal and chest wall nerves and can trace nerves from their dorsal nerve root ganglion (DRG) to the somatic chest/abdominal walls. Normal intercostal nerves on MRN are brightest at DRG and gradually fade in signal intensity with uniform size as they course posterior to anterior along the chest wall. Abnormal nerves demonstrate one or more of a combination of findings, such as persistent increased signal asymmetrical to the adjoining nerves, focal areas of increased signal, increasing size and focal neuroma like swelling.

There is a lack of systemic scientific literature on the role of MRN in the diagnostic assessment of intercostal neuralgia. Therefore, the aim of our study is to perform a cross-sectional analysis of patients presenting for MRN study with thoracic or abdominal pain and suspected intercostal neuralgia in our tertiary care institute and to evaluate the spectrum of nerves affected, the cost associated with MRN evaluation, their treatment course, and final outcomes. We hypothesized that MRN evaluation was a more cost-effective way of diagnosing intercostal neuralgia than the current state of using multiple imaging and invasive modalities, and that in cases with positive nerve abnormalities on MRN, selective nerve blocks and neurectomy aimed at the identified nerve(s) resulted in improved pain scores.

Methods and materials

This study was performed in compliance with health insurance portability and accountability act (HIPAA) regulations and was approved by the institutional review board of UT Southwestern Medical Center. Due to the retrospective nature of the study, informed consent was waived.

Patient population

A search of the electronic clinical database from our institution was performed for all adult patients (18–100 years old) who underwent an MRN for chest and/or abdominal wall pain from January 2012 to April 2019 seeking for the possibility of intercostal nerve pathology. The search terms included were “MR” and “intercostal”. The inclusion criteria were age 18–100 years, both genders, and clinical suspicion of intercostal nerve involvement, resulting in 313 cases. The exclusion criteria included whole body MR imaging performed for neurofibroma surveillance, as well as cases solely mentioning intercostal space, intercostal vessels, intercostal lymph nodes, intercostal lesions, intercostal muscles, and unrelated cases thus limiting the results to 33 cases. One case was excluded due to inability of the patient to complete the study and four additional cases were excluded for failing to follow-up with their physicians for a discussion of treatment options after MRN. Subsequently, the resulting 28 MRN studies and patient charts were systematically reviewed (Figure 1). The patients were imaged as standard of care for unclear cause of somatic chest wall pain.

Figure 1.

Flowchart depicting the initial recruitment process and data refining for the study cohort.

Data collection

The clinical data were collected through extensive chart review by two fellowship trained MSK radiologists and a medical student in consensus. Recorded data included patient demographics, inciting event, symptoms (pain or numbness) before MRN, symptom location, duration, and pain level. Other relevant data for work-up of the specific pain syndrome, such as type and number of imaging studies, procedures, etc. before MRN, and number of related clinical encounters was recorded. MRN findings and number of nerve involvements were extracted from the dedicated chest wall MRN reports, which had been created in the light of the clinical findings. These examinations and reports were again reviewed by two fellowship trained MSK radiologists and medical student in consensus. Since, we wanted to derive the impact of the these MRN reports on patient outcomes, the examinations were not re-read the second time around. Lastly, the post-MRN treatment plan and management were recorded from the electronic health records, including injection data (e.g. total number of injections and pain response duration) as well as any surgical data (e.g. degree of improvement and duration of follow-up). Patients who underwent injections were split into two groups (Group A vs Group B) to better visualize their course of management. Patients in Group A experienced fewer days of pain reduction during their second round of injections compared to their first round. Meanwhile patients in Group B experienced more days of pain reduction during their second round of injections compared to their first round. All information was recorded on a HIPPA-compliant and password protected Excel spreadsheet (Microsoft 2010).

Cost analysis

Medicare payment data were collected from the www.cms.gov database using the appropriate CPT codes for the respective procedures and office visits. An average of the facility and non-facility price per CPT code was calculated and then multiplied by the respective amount of usage per case. Individual cost analysis was performed for each case. Data from UT Southwestern Medical Center were also collected using the appropriate CPT code for the MRN used for each patient case. An average of all charges was calculated. If more than one CPT code was used for a case, an average charge was calculated, which was then used in the sum calculation for all cases.

MR neurography protocol and image analysis

A standard institutional MRN intercostal neuralgia protocol was performed in all cases (Table 1). The MRN was obtained on a 3-Tesla (3T) scanner (Achieva or Ingenia, Philips Healthcare, Best, Netherlands) using multiplanar 3D anatomical and fluid sensitive sequences. A torso XL multichannel coil was used in conjunction with spine coils. Thoracic spine imaging was part of the protocol to identify thoracic spine disease or radiculopathy, which can be a confounder in the diagnosis of intercostal neuralgia. The technologist marked the clinically affected site before the scanning.

Table 1.

3T MR neurography intercostal neuralgia imaging protocol

MR Sequence	2D/3D	TR	TE	Slice Thickness	Gap	Scan time
3 Plane Scout	2D	5.2 ms	1.37 ms	10 mm	10 mm
Coronal STIR TSE a	3D	3000 ms	78 ms	1.5 mm isotropic voxel	0	7 min
Ax SSFSE T2 FS b	2D	2000 ms	80 ms	5 mm	0.5 mm	4 min
Ax T 1W	2D	500 ms	8.8 ms	4 mm	0.5 mm	4.5 min
Ax T2 SPAIR (multivane) W/ RESP TRIGGERING c	2D	3582 ms	65 ms	4 mm	0.4 mm	4.5 min
Ax T2 through spine	2D	3000 ms	94 ms	5 mm	0.5 mm	4 min
Axial DTI	2D	11000 ms	87 ms	5 mm	0.0 mm	6 min

^aShort TI Inversion Recovery Turbo Spin Echo.

^bSagittal T ₂ weighted partial Fourier single shot fast spin echo w/ fat suppression.

^cAxial T2 spectral attenuated inversion-recovery megavoltage with respiratory triggering.

All MRN studies were independently interpreted by three fellowship-trained musculoskeletal radiologists with 2–10 years of experience with interpretation of MRN studies. Standard MRN reports included findings on lesions of soft tissue, spine, bones, muscles, peripheral nerves, masses, and viscera in the thoracoabdominal region. Specific MRN findings that confirmed clinical suspicion of intercostal neuralgia included increased T2 signal on fluid sensitive sequences and/or diffuse/focal thickening of the intercostal nerve(s) at the appropriate site of pain. Any regional scarring, neuroma or focal lesion were also reported.

An inter-reader reliability assessment was also performed after re-reads by a senior MRN reader to quantify the level of agreement on the MRN findings. An ordinal scale of 1 to 4 was created where “1” represented a negative MRN study, “2” represented one positive nerve identified, “3” represented two positive nerves identified, and “4” represented three positive nerves identified. The scale was applied to each case and the responses from each reader were recorded. A Cohen’s weighted κ was then calculated using the “Visualizing Categorical Data” software package within “R”, a software environment for statistical computing. ¹⁵

Ultrasound/CT-guided injection technique

The ultrasound and CT-guided nerve blocks were performed by musculoskeletal radiology fellows under the direct supervision of experienced fellowship trained musculoskeletal radiologists, by the radiologists themselves, or in one case, by a pain management specialist anesthesiologist. The patients were prepared and draped in usual sterile fashion. An optimal site for injection was then chosen and marked, after which a local anesthetic of lidocaine 1% without epinephrine was used to numb the area. Under CT or ultrasound guidance, a 22-gauge spinal needle was advanced to the target nerve site and optimal position between the inner and internal intercostal muscles was confirmed via imaging. A standard mixture of 2 ml of 1% lidocaine, 2 ml of 0.5% bupivacaine, and 1 ml of 4 mg ml⁻¹ dexamethasone was then injected around the target nerve and repeated as necessary for the other affected nerves. Post-procedural imaging was then used to confirm the appropriate localization of injection(s) and to evaluate for any possible complications, such as hematoma. After the injection(s), the patients attended a follow-up appointment to assess their pain improvement post-procedure. More than 50% improvement in pain lasting greater than 2 days was considered a positive response.

Nerve decompression surgical technique

Surgical neurectomies were performed by an experienced plastic surgeon specializing in neurolysis, nerve repairs, and neurectomy. The patients were prepared and draped in usual sterile fashion. An incision was made through the skin, subcutaneous tissue, fascia, and then partially through the serratus muscles until the intercostal muscles were encountered. The intercostal muscle was then dissected, and intercostal nerve(s) eventually found. Once identified, the nerve(s) was injected with 0.25% Marcaine with epinephrine, cauterized, and then cut for specimen. Subsequently, the nerve ending(s) was then buried into the surrounding muscle. A positive outcome was defined as a mean pain score of 2 or less with no recurrence of pain during follow-up.

Results

Patient demographics, clinical findings, and prior work-up

A total of 28 patients (age range, 18–82 years; mean age, 48.3 ± 18.0 years), including 21 female and 7 male patients (female-to-male ratio, 3.0) were included in the study. The most common location of pain was at the right upper abdomen (12/28; 43%). Frequency of anatomic distribution of symptoms are shown in Figure 2. Among various symptomatic sites, there was no patient with left lower quadrant abdominal pain.

Figure 2.

Schematic image showing the anatomic distribution and frequency of chest and abdominal wall pain in the study population.

The mean maximum pain level experienced was 7.4 ± 2.5 on a 1–10 visual analog scale (VAS). The mean duration of pain before MRN work-up was 36.9 ± 37.9 months. In addition, the patients reported a mean duration of 7.6 ± 9.5 months of worsening pain before undergoing MRN. All patients experienced pain or numbness (sensory symptoms) in the area of concern prior to MRN and had visited an average of 5 ± 2.8 physicians for the morbidity and its impact on patient quality of life. 20 (71%) patients had prior imaging studies for work-up of the pain including conventional MRI examinations, ultrasound, CT scans, and radiographs. Five (18%) patients underwent colonoscopy and/or esophagogastroduodenoscopy to exclude possible gastrointestinal etiology. Using centers for Medicare and Medicaid (CMS) data, the average Medicare payment per patient for prior work-up was calculated to be USD 1,006 ± 568 with a total sum of USD 28,175 added to healthcare costs for all 28 cases. Meanwhile from CMS data, the average cost per patient for an MR neurography was calculated to be USD 355 ± 105 with a total sum of USD 9,937 added to health-care costs for all 28 cases. Table 2 summarizes patient demographics, pain characteristics, and prior work-up done for these patients.

Table 2.

Demographics, pain, and other relevant clinical characteristics of the study population (n = 28)

Parameter	Value
Demographics
Gender (n = 28)
Male	7/28 (25%)
Female	21/28 (75%)
Age (y), mean ± SD	48.3 ± 18.0
Pain characteristics
Pain location
Right upper chest	1/28 (3.6%)
Right lower chest	3/28 (10.7%)
Left upper chest	3/28 (10.7%)
Left lower chest	2/28 (7.1%)
Right abdomen	16/28 (57.1%)
Left abdomen	3/28 (10.7%)
Epigastric	1/28 (3.6%)
Maximum pain level, mean ± SD (10-point scale)	7.4 ± 2.5
Duration of pain (mo), mean ± SD	36.9 ± 37.9
Duration of worsening of pain (mo), mean ± SD	7.6 ± 9.5
Management and health costs
Number of physicians visited for pain, mean ± SD	5 ± 2.8
Number of patients with prior imaging	20/28 (71%)
Number of patients with prior endoscopy/colonoscopy	5/28 (18%)
Medicare payment for prior work-up/patient, mean ± SD	$1006 ± 568
Medicare payment for prior work-up, sum (n = 28)	$28 175
Medicare payment for MRN, mean ± SD	$355 ± 105
Medicare payment for MRN, sum (n = 28)	$9937

MRN, MR neurography.

MR neurography findings

Among the 28 MRN studies with intercostal neuralgia, 19 MRN studies (68%) showed a positive evidence of intercostal neuropathy. 10 patients showed an abnormality of a single nerve while 9 patients showed multiple nerve abnormalities, with a total of 30 abnormal nerves identified among 19/28 patients. The ninth intercostal nerve was the most commonly affected nerve in this study with 7/19 patients (36.8%). Of the positive cases, MRN also detected post-surgical changes in chest/abdominal wall with scarring around peripheral nerves in 4/19 (21.1%) as well as neuroma formation in 7/19 cases (36.8%). The inter-reader reliability assessment resulted in a Cohen’s weighted κ value of 0.78. Figures 3 and 4 are representative images depicting the abnormal nerve findings identified with MRN.

Figure 3.

33-yo female s/p breast implant presents with left posterolateral chest wall tingling, numbness, and burning pain. Chest X-ray was unremarkable. (a) Axial T2 SPAIR image with an enlarged and hyperintense left ninth intercostal nerve (big arrow) with surrounding inflammation (small arrow). (b, d, e) 3D STIR TSE maximum intensity projection (MIP) images in multiple planes showing inflammatory changes (small arrows) in left anterolateral chest wall with thickened ninth (large arrow in b and e) worse than 10th (lower large arrow in D) intercostal nerve consistent with Sunderland Class III injuries. (c) Inverted grayscale diffusion tensor imaging of the chest confirms abnormally thickened 9th (upper arrow) nerves and to lesser degree 10th intercostal nerve (lower arrow) with effective vascular signal suppression. The patient responded positively to the perineural injections, which reduced her pain from an 8/10 to a 3/10. Surgery confirmed thickening of the 9th > 10 th nerves with an additional small neuroma of the 10th nerve (small arrow in e). MIP, maximum intensity projection; SPAIR, SPectral Attenuated Inversion Recovery; STIR TSE, short TI inversion recovery turbo spin echo.

Figure 4.

48-yo M with right sided posterolateral pain for at least 30 months prior to presentation who has seen four prior physicians. Previous ultrasound and MRI work-up were normal. (a) Axial reconstruction from 3D STIR TSE maximum intensity projection imaging show neuromas in continuity (arrows). (b) Inverted contrast diffusion tensor imaging confirms the Sunderland Class IV injury with neuromas in continuity of the ninth intercostal nerve. (c) Sagittal oblique reconstruction from 3D STIR TSE maximum intensity projection imaging show abnormal thickened nerve (large arrows) with neuromas in continuity (small arrows). The patient underwent surgery, which additionally identified large varicose vessels wrapped around the eighth and ninth intercostal nerves. STIR TSE, short TI inversion recovery turbo spin echo.

Management

Among the patients who were MRN positive for intercostal neuralgia, 4/19 received conservative managements including medications and physical therapy, 3/19 patients chose to address other health issues till the time of follow-up, 3/19 opted for only surgical management, 6/19 decided to receive only perineural injections, and 3/19 opted for both perineural injections followed by surgical management. Among the nine patients who were MRN negative for intercostal neuralgia, 6/9 patients decided for conservative management, one patient underwent incision and drainage for an MRN-diagnosed abscess, one patient received only perineural injections, and one patient opted for both perineural injections and surgical management. Table 3 exhibits a detailed breakdown of post-MRN management strategies.

Table 3.

Detailed management information after MRN

Parameter	Value
Positive MRN cases (n = 19)
Conservative management	4/19 (21.1%)
Opted to address other health issues	3/19 (15.8%)
Neurectomy	3/19 (15.8%)
Perineural injection	6/19 (31.6%)
Perineural injection + neurectomy	3/19 (15.8%)
Negative MRN cases (n = 9)
Conservative management	6/9 (66.7%)
Incision & drainage	1/9 (11.1%)
Perineural injection	1/9 (11.1%)
Perineural injection + neurectomy	1/9 (11.1%)

MRN, MR neurography.

Perineural injection and response

Perineural injection performed for 9 out of 19 patients (47%) with positive findings on MRN. Six out of the nine patients (67%) reported improvement on the first round of injections, lasting an average of 41.1 ± 83 days. There were two patients (22%) who reported no improvement and one patient (11%) was lost to follow-up. Five patients (56%) continued to receive a second round of injections, all of which reported pain improvement. However, 3/5 (Group A) reported a shorter duration of pain improvement on the second round of injections with an average of 19.7 ± 25 days. Notably, for these patients, they did not choose to pursue an additional third round of injections. On the other hand, the other 2/5 patients (Group B) reported a longer duration of pain improvement on the second round of injections with an average of 45 ± 55 days. These patients went on to have additional rounds of injections. The average total days of pain improvement from perineural injections was 150.2 ± 152.7 days. There were also two patients who decided to receive a perineural injection despite a negative MRN. Post-injection, one patient reported no improvement in symptoms and the other was lost to follow-up. See Figure 5 for a flow chart of patient management regarding perineural injections and their response.

Figure 5.

Flowchart showing clinical response rate to perineural injection and neurectomy in MRN (+) vs MRN (–) patients. *R = round #, *I = # of Improved patients, *NI = # of non-improved patients. MRN, MR neurography.

Surgical response

There were 7 out of 28 patients total (25%) who ultimately opted for a neurectomy performed by an experienced peripheral nerve surgeon. Among the six positive MRN patients who received a neurectomy, five patients reported improvement with an average follow-up period of 164.8 ± 181.5 days. The one MRN positive patient who did not obtain definite improvement experienced recurrence of pain 120 days later with a pain score of 6, after which a second neurectomy attempt was performed. That attempt proved successful in providing definite improvement with no recurrence of symptoms. On the other hand, one patient with a negative MRN underwent a neurectomy and obtained complete improvement in their pain. See Figure 5 for a flow chart of patient management in regard to neurectomy and their response.

Discussion

Chronic torso pain can be a complex syndrome with various culprits during different clinical scenarios. These devastating conditions could be attributed to visceral and somatic etiologies, each requiring different diagnostic work-up. Initial results of our study emphasize the importance of state-of-the-art diagnostic imaging that can provide fresh insights into the evaluation of such complex conditions. Underlying neurogenic etiology could be more precisely diagnosed in high detail by using advanced MR techniques tailored to evaluate the peripheral nerves, i.e. MRN. Among its advantages are the ability to suppress the venous and arterial signals and isotropic 3D evaluation that enhance the visualization of normal and abnormal nerve signals, therefore facilitating identification of nerve entrapment, perineural scarring, and neuroma changes. ^16–20 Positive MRN also allowed image-guided interventions and/or surgery with precise road mapping. Initial results also show that imaging and guided interventions ultimately resulted in improved outcomes and more appropriate expenditure of medical resources in this small-scale study. MRN was able to diagnose nerve pathology in 68% of cases (19/28) localizing 30 abnormal nerves.

The most common reported cause of chronic abdominal wall pain is anterior cutaneous nerve entrapment syndrome (ACNES). ^5,21 ACNES is defined as entrapment of the distal cutaneous branches of the lower thoracoabdominal intercostal nerves at the lateral border of the rectus abdominis muscle. ²² Other common causes include nerve entrapment in surgical scar, myofascial pain, diabetic neuropathy, thoracic spinal cord compression, and slipped rib syndrome. ²³ Chronic chest wall pain could also have numerous etiologies ²⁴ including intercostal neuralgia secondary to trauma, surgery, entrapment, or neuroma, isolated musculoskeletal entities such as costochondritis, rheumatic diseases such as fibromyalgia, and non-rheumatic conditions, such as neoplasms and sickle cell disease. There are multiple different etiologies of intercostal neuralgia. These include nerve entrapment, injury related neuritis or neuroma, persistent nerve irritation, pregnancy, nerve degeneration, muscle strain, and herpes zoster. ^1,25 Intercostal neuralgia can also be a post-op complication, most commonly seen with chronic chest wall pain after a thoracotomy or endoscopy, but it has also been reported in patients after breast and abdominal surgery, prior trauma, and infection. ^26–29 The pain could be dull or stabbing and is often present for several months or years. The pain is independent of food intake or bowel habits and is worsened with tight clothing, sneezing, coughing, laughing, and physical exercise. ^21,30 While MRN doesn’t show tiny cutaneous branches due to resolution limits, identifying main intercostal and subcostal nerves throughout their course along the chest and upper abdominal wall is not difficult. The neuropathy is conspicuous with enlarged, hypertense nerves with or without neuroma. Additional findings of perineural granulation tissue, scarring, abscess of space occupying lesion can be easily seen.

When clinicians are confronted with patient with chronic torso pain, the general assumption is that “pain is visceral until proven otherwise”. ² This approach often imposes significant financial burden to the patients and health care system through multiple clinical visits and costly diagnostic and invasive tests without helping to alleviate patient discomfort. This could be partly due to unfamiliarity of clinicians with various somatic causes of chest and abdominal wall pain. Accurate clinical distinction of visceral or somatic symptoms is not straightforward and has been addressed in the literature. ^31–34 In addition, neuropathologic pathways have been described explaining overlapping symptoms of visceral and somatic pains. Neural interaction at the level of posterior horn of the spinal cord between somatic and visceral pathways can also result in manifestation of visceral symptoms related to autonomic stimulation, such as anorexia, nausea and vomiting, in patients with somatic pathologies. ^35,36

John Carnett first described intercostal neuralgia as a cause for chronic abdominal wall pain in a hallmark paper dates to 1926. ³⁷ Positive Carnett’s sign is described as exacerbation of focal pain over the area of maximal abdominal tenderness when straight leg raising, or chin-to-chest position tenses the abdominal wall. This clinical test has reportedly high positive and negative predictive values in differentiating somatic from visceral sources of abdominal pain ³⁸ and is 78% sensitive and 88% specific. ³⁹ If Carnett’s sign is positive, a trigger-point injection of anesthetics and steroids is often recommended as a useful diagnostic and therapeutic exercise. ^40–43 It is important to note that high placebo effect with injections can exist. ⁴⁴ However, clinical examination is often limited in identification of the underlying etiology for somatic pain syndromes. In addition, these painful conditions could have referral nature with origin in distinct anatomic location from maximal pain sensation. The complexity of the underlying etiologies for chronic torso pain, subjectivity and lack of precision of clinical evaluation, and the mental and financial burden of such devastating conditions warrant utilization of advanced diagnostic modalities for earlier diagnosis and more accurate treatment planning.

MRN is a dedicated MR imaging technique that uses state-of-the-art pulse sequences and imaging protocols to visualize the peripheral nerves. The introduction of 3D imaging pulse sequences and improvement in diffusion MR techniques have considerably evolved evaluation of peripheral nerves. ^45,46 State-of-the-art advances in MR technology including dedicated multi channel radiofrequency surface coils have made it feasible to obtain high resolution, high contrast MRN imaging to delineate the small peripheral nerves, their abnormalities and related features.

MRN was able to identify nerve pathology and accurate location of nerve abnormality in 68% (19/28) of patients in our patient cohort. Different forms of treatment have been used for chronic abdominal and chest wall pain. The most common are narcotic medications, trigger-point injections, intercostal nerve blocks, neurectomy, transcutaneous nerve stimulation, and celiac plexus neurolysis. ^2,3,47 Among these, injections and nerve blocks are less invasive without risk of opioid abuse. There was significant improvement (67%) in symptoms with perineural injection (6/9 patients). Most of patients required more than one rounds on perineural injection for symptomatic relief (5/6 with 2 round and 2/6 with three rounds). Three patients in this group (3/9, 33%) did not respond well to perineural injection, which subsequently became symptom-free following neurectomy. Three patients with positive findings on MRN decided to pursue neurectomy without preceding perineural injection and all demonstrated symptom relief with favorable outcome. MRN was able to identify culprit pathology (soft tissue abscess) in a patient with normal nerve findings. These findings emphasized significant role of state-of-the-art MRN technique for diagnosis and management planning of patients with chronic chest and abdominal wall pain. Ten (10/28, 36%) patients (four in MRN positive and six in MRN negative groups) pursued conservative pain management without intervention. Perineural injection was performed for two patients despite negative results on MRN and it did not result in symptomatic relief. Surgical management was often successful with a positive outcome in 6 out of 7 operated cases (85.7%).

Patients with chronic chest and abdominal wall pain often undergo a variety of usually unnecessary diagnostic examinations. Studies have reported a wide range of cost occurrences before the diagnosis of such pain syndromes, ranging from USD 700 to 6,727 USD per patient. ^8,48,49 The average Medicare payment per patient for prior work-up was calculated to be 1006 ± 568 per patient in our cohort. Our result is in line with a report by Costanza et al ⁸ , which reported an average cost of USD 1133 per patient for diagnosis of chronic abdominal wall pain. Glissen Brown et al ⁵ have reported that “the diagnosis of chronic abdominal wall pain is clinical rarely requiring supplemental lab work or imaging”. However, they have highlighted a group of patients under the category of “chronic abdominal pain of uncertain etiology”. Our cohort fits into this category and included patients with chronic pain (36.9 ± 37.9 months of pain) of unknown source referred by gastroenterologists or plastic surgeons in a tertiary care center. MRN was successful to find the pathology (post-surgical perineural scarring, neuroma, or involvement of single or multiple nerves) in 68% (19/28) of our cohort. In addition, MRN depicted precise location of the pathology which is of utmost importance for percutaneous injection or surgical neurectomy planning. The costs were also about one-third with MRN for diagnostic purposes with less financial and psychological harm.

This study does have limitations of a retrospective study, wherein the methodology was not prospectively controlled. We are limited by our sample size of 28 subjects, especially in subgroups of positive and negative MRN findings. This small sample size, especially for the negative MRN subgroup, could be due to intercostal neuralgia being a diagnosis of exclusion and thus hesitation to order a MRN as an initial study. Providers may also be unaware of MRN as an option for assessing neuropathic pain due to its novelty and lack of studies like ours, and as such are less likely to order such a study. In addition, some patients initially included in the study may have sought for medical care for chest and/or abdominal wall pain outside of our institution, and thus their treatment response after MRN was not possible to be assessed. The study also lacks from comparison to a control group in terms of efficacy of perineural injection and neurectomy. Chest and abdominal pain syndrome could have multiple simultaneous concomitant reasons. Investigators were not able to eliminate these cofounders from their evaluation. Another limitation is that consistent pathological confirmation of the MRN imaging findings was not available and thus, unable to be assessed in this study. Therefore, there is a possibility of confirmation bias. However, our inter-reader reliability assessment demonstrated a high Cohen’s weighted κ value of 0.78, which is interpreted as “excellent agreement” by Cicchetti. ⁵⁰ Such high agreement on the MRN findings decreases the likelihood of confirmation bias. The study also lacks from a qualitative imaging assessment of pathology. Measurement of changes in nerve signal intensity and thickness can be difficult due to the small size of the intercostal nerves. However, acquisition of 3D imaging with zero gap and maximum intensity projection reconstructions created along the plane of the chest and the nerves likely mitigated such partial voluming effects. In regards to these limitations, the topic would benefit from future prospective studies having a larger sample size, a longitudinal correlation of treatment response with statistical evaluation, and consistent pathological confirmation of neuropathy using surgical specimens. These future steps seem necessary to highlight the clinical efficacy of MRN in the domain of chronic chest and abdominal wall syndrome.

Despite limitations regarding the quantity of patients undergoing MRN, this study provides early data supporting the usefulness of MRN as an imaging technique to directly visualize nerve pathology and in turn, identify candidates who will most likely benefit from perineural injections and/or neurectomy. As such, there is promise that the current diagnostic and management strategy of intercostal neuralgia can be improved with MRN through its ability for prompt nerve pathology identification. However, further studies with larger sample size as well as a comparison to intercostal neuralgia diagnosed without MRN will need to be performed to further clarify MRN’s impact on intercostal neuralgia. Regardless, for intercostal neuralgia, unnecessary clinical visits, diagnostics, and therapeutic procedures have been the norm for many years and there needs to be a movement towards more appropriate utilization of health-care resources.

In summary, application of state-of-the-art MRN is promising for more accurate diagnosis and treatment planning of patients with chronic chest and abdominal wall pain. This can prevent unnecessary clinical visits and diagnostic and therapeutic procedures as well as more appropriate utilization of health-care resources, which could ultimately result in reduction in health-care expenditure with improved management strategy of such complex patients.