Origin, Branching, and Communications of the Intercostobrachial Nerve: a Meta-Analysis with Implications for Mastectomy and Axillary Lymph Node Dissection in Breast Cancer

The intercostobrachial nerve (ICBN), which usually originates from the lateral cutaneous branch of the second intercostal nerve, innervates areas of the axilla, lateral chest, and medial arm. It is at risk for injury during operative procedures that are often used in the management of breast cancer and such injury has been associated with postoperative sensory loss and neuropathic pain, decreasing the quality of life. PubMed, Excerpta Medica Database (EMBASE), ScienceDirect, Google Scholar, China National Knowledge Infrastructure (CNKI), Scientific Electronic Library Online (SciELO), Biosciences Information Service (BIOSIS), and Web of Science were searched comprehensively. Data concerning the prevalence, branching, origin and communications of the ICBN were extracted and pooled into a meta-analysis. A total of 16 studies (1,567 axillas) reported data indicating that the ICBN was present in 98.4% of person. It most often (90.6%) originated from fibers at the T2 spinal level and commonly coursed in two branching patterns: as a single trunk in 47.0% of cases and as a bifurcating pattern in 42.2%. In the latter cases, the bifurcation was usually unequal (63.4%). Additionally, the ICBN presented with anastomosing communication to the brachial plexus in 41.3% of cases. The ICBN is a prevalent and variable structure at significant risk for injury during operative procedures of the axilla. In view of the postoperative pain and paresthesia experienced by patients following injury, surgeons need to exercise caution and aim to preserve the ICBN when possible. Ultimately, careful dissection and knowledge of ICBN anatomy could allow postoperative complications to be reduced and patient's quality of life increased.


Introduction And Background
The intercostobrachial nerve (ICBN) is a nerve classically originating from the lateral cutaneous branch of the second intercostal nerve [1]. The ICBN functions to innervate portions of the axilla, tail of the breast, lateral chest wall and medial side of the arm [2][3]. The ICBN, in its historically portrayed course, exits the second intercostal space and traverses the axilla to terminally branch in the region of the medial arm [1]. The proximity and course of the ICBN in relation to the axilla poses a danger of iatrogenic injury resulting from common procedures such as axillary lymph node dissection (ALND), sentinel lymph node biopsy (SLNB) and mastectomy [2,4].
Breast cancer is the most common malignancy in women worldwide, affecting nearly one in eight in both the United States and Europe [1,[5][6]. Procedures such as ALND, SLNB, and/or mastectomy are utilized in a significant portion of breast cancer cases for purposes ranging from diagnosis, staging, to resection [1,[7][8]. The ICBN is the most commonly injured nerve during mastectomy and is believed to be implicated in both persistent pain after breast cancer treatment (PPBCT) and permanent loss of sensory function in the region supplied [1][2]9].
It has been estimated that between 10-60% of breast cancer survivors who have undergone operative procedures, experience PPBCT, and it is believed that this predominantly neuropathic pain contributes significantly to a reduced quality of life following breast cancer treatment [1,[9][10]. Aside from dysesthesia, paresthesia of the ICBN, sensory distribution is also very common [11]. It has been shown that preservation of the ICBN provides for a clear reduction in post-procedural paresthesia and improves the quality of life [1,11]. Surgeons should exercise precision during dissections of the axillary region as early branching and anastomosing fibers between the ICBN and brachial plexus can easily be overlooked [1,[12][13].
ICBN anatomy has been studied in detail in the past two decades, yet on a very limited scale. The ICBN's structure is highly variable in nature with numerous origins, branching patterns and communicating branches [1, 4-13-17]. Determining the true prevalence and nature of the variant patterns of ICBN branching and origin will serve surgeons to better preserve these neural structures. A surgeon's knowledge of these structures is vital to their safeguarding during operative breast cancer management. The clinical importance of the ICBN necessitates further investigations in the future, because despite the development of detailed anatomical investigations that are becoming more widespread, the ICBN remains an overlooked structure.
The aim of this study was to utilize all available data to date to provide a comprehensive evidence-based appraisal of the prevalence and varying patterns of ICBN origin and branching. It is our intention, that this information be utilized to help decrease the morbidity and risk of postoperative complications of breast cancer treatment.

Search Strategy
Identification of articles for inclusion in the meta-analysis was performed with searches through December 2016 in the following databases: PubMed, EMBASE, ScienceDirect, Google Scholar, China National Knowledge Infrastructure (CNKI), SciELO, BIOSIS, and Web of Science. The search strategy executed in PubMed is presented as follows: (Intercostobrachial nerve[title/abstract]) or nervus intercostobrachialis[title/abstract] No date or language restrictions were applied. Identification of additional studies eligible for the meta-analysis was performed through searching the references of all included articles. Throughout the meta-analysis, adherence to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines were exercised (See Appendix - Table 5) [18].

Eligibility Assessment
Eligibility for inclusion in the meta-analysis was assessed by two independent reviewers. All intraoperative or cadaveric studies that reported extractable prevalence data of ICBN origin or branching were included. The exclusion criteria included: case reports, case series, letters to the edito or conference abstracts. All studies published in languages not fluently spoken by any of the authors were translated by medical professionals fluent in both English and the language of the original manuscript. Disagreements between reviewers arising during the eligibility assessment process were resolved by discussion and consensus.

Data Extraction
Data from included studies was independently extracted by three reviewers. Extracted data included year, country, sample size (number of patients and number of axillas), prevalence of ICBN, ICBN origin, branching pattern of ICBN and the prevalence of communication with the brachial plexus. In the event of any discrepancies in the data, authors of the original were contacted for clarification when possible.

Statistical Analysis
The single-categorical and multi-categorical pooled prevalence estimates were calculated using MetaXL version 2.0 by EpiGear Pty Ltd. (Wilston, Queensland, Australia) [19]. A random effects model was used for all statistical analyses. Heterogeneity was assessed by both the Chi 2 test and the I 2 statistic. For the Chi 2 test, a p-value of < 0.10 for Cochran's Q served as an indicator of significant heterogeneity among the studies analyzed [20]. The results of the I 2 statistic were interpreted as follows: zero percent to 40% might not be important; 30%-60% could indicate moderate heterogeneity; 50% -90% could indicate substantial heterogeneity; and 75% -100% could represent considerable heterogeneity [20].
Subgroup analysis was performed on the basis of type of study (cadaveric vs. operative) and geographical origin of the study. Statistically significant differences between analyzed groups were determined by their confidence intervals. If the confidence intervals of any two rates overlapped, the differences were regarded as statistically insignificant [19]. Lastly, sensitivity was assessed by a separate analysis of studies with sample sizes greater than 100 to probe further for potential sources of heterogeneity.

Study Identification
The flow of studies through the meta-analysis is presented in Figure 1. The search of the major electronic databases identified an initial 509 articles, with an additional one article identified in the search of the references of included studies. A total of 102 articles were assessed for eligibility using full-texts, of which 86 were excluded and 16 were included into the metaanalysis.

Characteristics of Included Studies
The characteristics of the studies that have been included in the meta-analysis are summarized in Table 1. Sixteen studies (1,567 axillas) [1, 3-4, 13,17, 21,28] were included in the meta-analysis. Five studies were performed on cadaveric specimens [13,15,22,25,27], 10 were performed intraoperatively [1,3-4, 14, 16-17, 21, 24, 26, 28], and one study [23] examined both cadavers and operative subjects. The studies ranged in year from 1999 to 2016 and had a vast geographic breakdown with 10 from Asia, three from Europe, and one each from North America, South America, and Oceania.   (Figure 2). No significant differences in prevalence were noted during geographic or study type subgroup analysis. Sensitivity analysis also proved to not reveal any notable differences. Data on the prevalence of the ICBN is presented in Table 2.   . To further clarify, combined origin was listed as such when two separate roots were observed unifying into one common ICBN. Subgroup and sensitivity analysis did not yield any significant findings or differences between groups. Data regarding the origin of the ICBN is presented in Table 3.

Discussion
The ICBN is a nerve of significant importance in surgical procedures involving the axillary region. Damage to the ICBN is associated with significant morbidity and is believed to be involved with postoperative pain and sensory loss [1,[9][10].
Our analysis demonstrated the overall prevalence of the ICBN to be 98.4% and that it most often exists as a single trunk (47.0%) originating from the T2 vertebral level (90.6%). Subgroup analysis revealed little variation in the results aside from that of ICBN branching. Branching subgroups differed largely based on the modality of the study (cadaveric vs. operative). When examining the statistics for bifurcation alone, it is notable that ICBNs examined in cadavers (65.3%) had more than twice the rate of bifurcation than those assessed intraoperatively (30.2%). We posit two concepts from these differences. The first being that due to the limited field of view intraoperatively, as well as the inability to freely dissect tissue without consequences, may limit the branches which are identified in-vivo. It should be noted that these differences may help explain why nerves which are successfully identified and protected can still present with postoperative complications [1]. It was noted in Kumar (2016) that six months following the procedure, 20% of patients who had successful ICBN preservation still experienced numbness and paresthesia [1]. However, in addition to a bifurcating ICBN, this may also be due to stretching of the nerve by retractors or other intraoperative stress placed on the nerve fibers. The secondary factor contributing to the cadaveric-intraoperative discord may be that nerves that do bifurcate, have a propensity to do so in an unequal fashion (63.4% unequal versus 36.6% equal bifurcation). As such, surgeons may easily identify the larger trunk but neglect to identify the smaller branch of the nerve, placing it at a higher risk of operative injury.
Preservation of the ICBN is very important and should be factored into operative plans. There are cases where ICBN conservancy will not be possible, but in those where it is possible, it was associated with an insignificant increase in operative time of approximately five minutes [1].
The neural fibers travelling within the ICBN appear to vary, as there are cases of known ICBN loss where patients have no postoperative neural deficits [1]. As it is difficult to determine preoperatively the role of the ICBN in a patient, a vital sensory function of the nerve should be assumed. It was also discovered in our analysis that 41.3% of ICBNs contributed anastomosing branches to the brachial plexus, namely the medial cord or medial brachial cutaneous nerve. This lends some explanation to the unpredictability of deficits that can present in patients [13]. Losses in sensation or pain have been noted in locations ranging from the medial side of the proximal arm, axillary region, to the anterior chest wall [1,17].
Various techniques of dissection have been proposed throughout the history of ICBN examination. Some techniques advocate that a superficial blunt dissection should be conducted from the site of axillary incision towards the lateral edge of the pectoralis muscle [12]. The ICBN can be identified consistently in an anterior position during exposure of the long thoracic and thoracodorsal nerves [13,[29][30]. Extra care should be particularly exercised in the area of the second intercostal space where ICBN origin is most likely (90.6%). Surgeons should however, consider than many ICBNs do receive fibers from either the first or third intercostal nerves as well. Some also note that dissection could propagate inferiorly from an axillary incision, traveling from the axillary vessels, layer-by-layer until positive nerve identification [4]. It is also noted that in the area of the inferior margin of the pectoralis minor muscle and dissection deep to the pectoralis minor, should be avoided if possible due to risks of ICBN injury [31].
Operative procedures of the axilla are often associated with varying stages of breast cancer management. The ICBN can be damaged from any number of reasons ranging from traction to transection [32]. Additionally, damaged nerves may also develop neuromas that can further complicate the clinical picture [1,32]. Surgeons in some cases must make decisions that have consequences in order to get the proper margins in oncologic surgery. It should, however, be standard to exercise protective practices towards the ICBN when possible. ICBN neuralgia and post-mastectomy pain syndrome can be successfully managed with loco-regional anesthetic techniques, however, overall reduction in incidence should be the primary goal [4,32].
Our meta-analysis was limited by a number of factors involved in inconsistent reporting and small sample sizes. Some studies presented origins from the first intercostal nerve or third intercostal nerve as separate data. It was unclear to whether these origins were duplications or simply contributing fibers and thus had to be excluded from analyses where this information could have skewed results. Additionally, high heterogeneity persisted despite our subgroup analyses, thus we postulate this may have resulted from the intrinsically variable nature of the ICBN. This high level of heterogeneity was sustained in defiance of the fact that pooled prevalences were largely consistent between groups, pointing further towards the notion that there is indeed high anatomic variability. There was an overall lack of reporting by studies on factors such as gender and laterality, which could have provided for additional subgroup analyses. The overwhelming majority of studies were performed on women, however, Loukas (2006) noted no differences between genders [15]. Limited data from regions such as Africa, North America, Oceania and South America precluded analyses of other regions aside from Europe and Asia. Detailed data was absent from many studies regarding branching site or details on the symmetry of the post-division branches.
Future studies need to be performed to further elucidate the behavior of the terminal branching of the ICBN as well as its anastomotic behavior with the brachial plexus. There is also the possibility to investigate the use of landmarks or relationships with adjacent structures as a means for intraoperative nerve identification. Structures such as the lateral thoracic vein discussed in O'Rourke (1999) could be a potential candidate [13]. Finally, histologic studies should be performed to examine for possibilities of microscopic branching and further identify the types of neural fibers running within the nerve. Investigating this may help explain cases of postoperative neuropathy where nerves were successfully spared intraoperatively. Supplementation with future research will provide valuable information for surgeons and ideally result in more positive outcomes for patients undergoing procedures in the region.

Conclusions
The ICBN is a prevalent anatomical structure with significant variability in origin and branching pattern. It is at considerable risk during operative procedures in the axilla, especially those involved in breast cancer management. Injury to ICBN fibers has been associated with post-procedural pain, paresthesia, and ultimately reduced the quality of life. Surgeons need to exercise caution and should adopt measures to protect the ICBN, its branches, and anastomosing fibers if possible. The information presented in this study, accompanied by careful surgical practice, could help to reduce significant complications during mastectomy and axillary dissection. consider their relevance to key groups (e.g., healthcare providers, users, and policy makers). 10-13

Limitations 25
Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).

Conclusions 26
Provide a general interpretation of the results in the context of other evidence, and implications for future research. 13

Funding 27
Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review. 13