12 Rehabilitation After Neck Dissection Abstract Rehabilitation after neck dissection is perceived as important, but there has been little research on specific regimens or protocols to help patients recover after neck dissection. This chapter evaluates the types of impairment following neck dissection, treatment factors that influence impairment, the assessment tools used to evaluate function, and optimal rehabilitation following neck dissection. Emphasis is placed on evidence regarding the impact of the extent and type of treatment on shoulder impairment. The literature is reviewed to determine the evidence supporting different rehabilitation modalities for both shoulder impairment and lymphedema. Best practice suggestions are provided based on synthesis of the content of the chapter in the conclusion section. Keywords: neck dissection, rehabilitation, shoulder impairment, lymphedema, pain, dysphagia, head and neck cancer Neck dissection is a surgical procedure commonly performed to improve regional control of head and neck malignancy of the upper aerodigestive tract, thyroid, parotid, or skin. The extent of neck dissection is determined by the primary tumor and the presence and location of regional metastasis. Most patients undergoing selective neck dissection (SND) experience relatively little disability. When neck dissection is more extensive, and when there is injury to the accessory nerve or when multiple treatment modalities are employed, there is an increased likelihood of impairment. The impairments include loss of skin sensation, paresthesia, loss of depressor anguli oris function, decreased range of motion of the neck and shoulder, neck pain, neck stiffness, and loss of strength for lifting of objects. Recent research has evaluated psychosocial and psychological sequelae in addition to functional impairments of neck dissection that warrant consideration in proper rehabilitation. This chapter describes the impairment following neck dissection, the treatment factors that influence impairment, the assessment tools used to evaluate function, and optimal rehabilitation following neck dissection. Shoulder impairment following radical neck dissection (RND) was first described by Ewing and Martin in 1952.1 The prevalence of impairment following neck dissection, predominately manifested by shoulder and upper limb impairment, ranges between 18 and 77%.2 Prevalence is difficult to report on because of the different types of neck dissection, the different primary site treatments associated with neck dissection, and the number of different measures that are used to assess neck dissection–related impairment. There are a number of patient factors that can contribute to the difficulty of neck dissection and potential complications postoperatively. Factors that limit surgical exposure such as morbid obesity and decreased range of motion of the neck increase likelihood of damage to the neural and vascular structures in the neck. Patient comorbidities such as smoking, alcohol consumption, diabetes, and coagulopathy can increase the likelihood of a wound complication that in turn can increase postoperative scarring and mobility. Patients with a history of atherosclerosis have an increased likelihood of stroke. The late effects of neck dissection include impaired neck and shoulder mobility and strength, cervical paresthesia, pain, lymphedema, dysphagia, psychosocial issues,3 and impaired quality of life.4 Evaluation of patients who underwent neck dissection and comparing to patients without neck dissection revealed impairments in intelligibility of speech, health-related quality of life, and decreased employment.3 Similar results were obtained using the University of Washington-QOL and Functional Assessment of Cancer Therapy Head and Neck questionnaires in patients who did and did not undergo neck dissection. Five years after treatment, patients who underwent neck dissection had worse scores with regard to aesthetics, willingness to eat in public, decreased levels of activity, and decreased involvement with recreation or entertainment than patients who did not undergo neck dissection.4 There are critical structures in each level of the neck that, when dissected, can contribute to impairment following neck dissection ( Component of level I include the following: level Ia, the submental nodes, extending from the mandible anteriorly, bordered by the anterior belly of the digastric bilaterally, and the hyoid inferiorly. Removal of fibrofatty tissue from this area can lead to a cosmetic contour deformity. The area can be not only depressed, but also excessively full due to lymphedema. It is generally thought that there is no treatment for these sequelae. Level Ib, which is bordered by the mandible superiorly, and the digastric muscle inferiorly, contains multiple neurovascular structures. The most notable structures include the hypoglossal and lingual nerves. Injury to the hypoglossal nerve, which supplies ipsilateral motor innervation to the tongue, can lead to dysphagia and dysarthria. The ansa hypoglossi is a branch of C1 and travels with the hypoglossal nerve to innervate the thyrohyoid and the geniohyoid. The nerve branch to the thyrohyoid (which passes to level III) is frequently cut during neck dissection, but the impact of this transection is not well understood. The lingual nerve supplies sensory innervation to the ipsilateral tongue. Dysgeusia is very uncommon after neck dissection. Table 12.1 Contributors to neck and shoulder disability based on level of neck dissection
12.1 Introduction
12.1.1 Prevalence of Impairment Following Neck Dissection
12.1.2 General Factors That Affect Functional Outcome after Neck Dissection
12.1.3 Late Effects of Neck Dissection
12.2 Components of Impairment after Neck Dissection
Table 12.1).
12.2.1 Level I
Anatomical level | Structure affected | Impairment/disability |
Level Ia | Fibrofatty tissue | Mild cosmetic deformity |
Level Ib | Hypoglossal N Lingual N Marginal mandibular N | Ipsilateral tongue hemiplegia, dysphagia, dysarthria Ipsilateral tongue paresthesia, dysphagia, dysgeusia, dysarthria Paralysis of lower lip depressor, cosmetic deformity, lower lip trauma |
Level IIa | Spinal accessory N Hypoglossal N Great auricular N | Shoulder and neck ROM and strength Ipsilateral tongue hemiplegia, dysphagia, dysarthria Ipsilateral pinna paresthesia |
Level IIb | Spinal accessory N | Shoulder and neck ROM and strength |
Level III | Phrenic N Ansa cervicalis N | Hemi diaphragm paralysis/DOE, pneumonia Hyolaryngeal elevation |
Level IV | Phrenic N Thoracic duct | Hemidiaphragm paralysis/DOE, pneumonia Chyle leak |
Levels II–IV | Jugular vein Vagus N Cervical rootlets Sympathetic trunk Carotid artery | Lymphedema Ipsilateral vocal cord paralysis, dysphonia, aspiration Cervical paresthesia Horner’s syndromea TIA, stroke |
Level V | Spinal accessory Nb Brachial plexus Cervical rootlets | Shoulder and neck ROM and strength Hand and arm paresthesias and weakness, hand or arm paralysis, severe pain Cervical paresthesia |
Abbreviations: DOE, dyspnea on exertion; N, nerve; ROM, range of motion; TIA, transient ischemic attack. b Increased shoulder impairment with dissection of this level. | ||
The marginal mandibular nerve is a branch of the facial nerve and it innervates the depressor anguli oris muscle, which depresses the ipsilateral corner of the mouth. The nerve courses inferior to the mandible within the fascia overlying the submandibular gland and can be injured during dissection of this level. Rates of injury are reported to be up to 23% in one observational study of 66 patients.5 The main complaint of patients following injury to the nerve is cosmetic deformity manifested by smile asymmetry. Patients may also complain of biting their lower lip when eating and some difficulties with oral competence.
12.2.2 Level II
Level II extends from the skull base superiorly to the hyoid inferiorly. The medial border is the deep cervical fascia overlying the paravertebral muscles and the levator scapulae. Level II is divided into level IIa that is superior to the spinal accessory nerve (SAN; cranial nerve [CN] XI) and with level IIb that is below CN XI.
CN XI supplies motor innervation to the sternocleidomastoid muscle (SCM). The motor innervation of the upper trapezius is supplied by variable contributions of CN XI and the cervical plexus. The sacrifice of CN XI results in shoulder impairment in 60 to 80% of patients.6 Even when intact, the extent of CN XI dissection is correlated with the degree of shoulder impairment.7,8 Deficits in shoulder function from dissection or sacrifice include reduced abduction and decreased range of motion, scapular winging, scapular droop, neck stiffness, and neck pain.
12.2.3 Level III
The boundaries of level III include the hyoid superiorly to the inferior border of the cricoid inferiorly, and the sternohyoid and deep cervical fascia over the paravertebral muscles medially. The largest segment of the ansa cervicalis provides motor innervation to the infrahyoid strap muscles to induce hyolaryngeal stabilization and depression. The contribution of these muscles is not well understood. The infrahyoid strap muscles are a counterbalance to the suprahyoid muscles. Contraction of the suprahyoid musculature produces hyoid and laryngeal elevation and causes the base of tongue to cover the laryngeal inlet and fold over the epiglottis (epiglottic inversion). The infrahyoid strap muscles provide a counterbalance to this swallowing action. Transection or resection of the ansa branches or the infrahyoid strap muscles do not seem to result in swallowing or airway impairment. If the patient undergoes multimodality (radiation) treatment that includes extensive resection including the SCM and the infrahyoid musculature, decreased hyoid elevation is observed during swallowing that can lead to increased incidence of aspiration.
12.2.4 Level IV
Level IV is bounded by the cricoid superiorly and the clavicle inferiorly. The medial border is the deep cervical fascia over the paravertebral muscles. The posterior border is aligned with the posterior border of the SCM. The structure most commonly injured in level IV is the thoracic duct, which is on the left, although there are large lymphatics found on the right. Injury to the lymphatics in level IV can lead to a chylocele that, in some cases, does not spontaneously resolve and may need treatment to resolve electrolyte loss that can lead to electrolyte abnormalities.
12.2.5 Common Structures in Levels II to IV
Structures common to levels II to IV include the internal jugular vein (IJV), carotid artery, vagus nerve, sympathetic trunk, and the SCM. Sacrifice of a single IJV contributes to lymphedema of the neck that can result in increased neck stiffness and an impaired cosmetic appearance. If both IJVs are transected, life-threatening cerebral edema can result. If both are transected acutely, a bypass should be performed. Stroke can result from manipulation of the carotid vessels; however, there appears to be no long-term sequelae from ligation of the external carotid artery. The vagus nerve provides innervation important for swallowing and it innervates the ipsilateral vocal cord; therefore, injury to the vagus can result in dysphagia, dysphonia, and aspiration. Finally, the sympathetic trunk courses within the carotid sheath, and injury can result in Horner’s syndrome of ptosis, meiosis, and anhydrosis.
12.2.6 Level V
The posterior triangle of the neck is bounded by the trapezius posteriorly, the clavicle inferiorly, and, for surgeons, the cervical rootlets anteriorly. It is divided into sublevels Va (superior) and Vb (inferior) by the plane of the inferior border of the cricoid. Level V contains CN XI. CN XI picks up branches of C2 as CN XI exits the SCM. The branches of C2 can variably provide motor innervation to the trapezius. Dissection of CN XI results in worse shoulder function in most patients.7,8,9 Dissection or sacrifice, of cervical rootlets, can lead to neck, earlobe, or upper chest paresthesia.
12.2.7 Level VI
Level VI is bordered superiorly by the hyoid, inferiorly by the innominate artery, and laterally by the carotid arteries. The deep boundary is the visceral fascia around the thyroid and larynx. Critical structures include the recurrent laryngeal nerves and the parathyroid glands. Dissection of the recurrent laryngeal nerve or the external branch of the superior laryngeal nerve can lead to weakness and changes in voice production. Dissection of the inferior thyroid artery of the parathyroid glands can lead to hypocalcemia.
12.3 Treatment Factors
If a neck dissection includes more levels, there will be more impairment—particularly if level V is dissected. It is important to note the both radiation and radiation with chemotherapy also contribute to impairment after neck dissection.
12.3.1 Types of Neck Dissection
The underlying disease determines the extent of neck dissection, and it should be the foremost determinate as to the type of neck dissection performed. The classification of neck dissection includes RND, modified RND (MRND), and SND.
RND involves removal of the ipsilateral cervical lymph nodes from levels I to V with sacrifice of the SCM, CN XI, and the IJV. The other neck dissections are variations of the RND with regard to levels of dissection and preservation of nonlymphatic structures. MRND also involves removal of ipsilateral cervical lymph nodes in levels I to V, but with preservation of one or more of the aforementioned nonlymphatic structures (SCM, CN XI, IJV). Functional neck dissection is a term that is no longer used that describes preservation of the SCM, CN XI, IJV, part of the cervical plexus, or one or more levels of the neck. SND involves removal of one or more nodal levels with preservation of nonlymphatic structures.
12.3.2 Extent of Neck Dissection
RND is associated with the worst functional outcome by virtue of sacrifice of vital structures for shoulder and neck function as well as cosmetic deformity with removal of the SCM. A retrospective review that assessed 224 patients who underwent 308 neck dissections evaluated pain, impaired shoulder function by measuring shoulder drop, reach above and arm abduction, increased neck and shoulder stiffness, and increased neck constriction for patients undergoing radical, MRND or SND. Sacrifice of CN XI was associated with worse outcome across all measures. If level V was not dissected, then patients had better outcomes across all measures. Cosmetic outcome was associated with the preservation of the SCM.10
Multiple studies validate the association of dissection of CN XI and increased impairment with different outcome measures. A retrospective comparison of SND levels II to IV (n = 20) with SND levels II to V (n = 20) showed that level V dissection was associated with impaired shoulder muscle strength, shoulder range of motion limitation, shoulder droop, protraction, and flaring, and decreased electromyographic potentials.7 Impairment in strength and function was predominately reported as mild. Similarly, retrospective review compared 15 patients who underwent MRND with preservation of the SCM, CN XI, and IJV with 17 patients who underwent SND levels II to IV. This study demonstrated a trend toward increased pain, increased disability as assessed by the Shoulder Pain Disability Index (SPDI), and decreased range of motion as measured by goniometry 6 months postoperatively in the MRND group despite sparing SCM, CN XI, and IJV.11 This finding is further corroborated in a retrospective review of 121 SND and 46 SCM and CN XI–sparing MRND where impaired shoulder function was found using the Neck Dissection Impairment Index (NDII) in the patients who received MRND.8
12.3.3 Dissection of Sublevel IIb
There is controversy in the field regarding the utility of sublevel IIb dissection. Multiple papers have evaluated patients who have a clinical and radiological N0 classification neck and showed the incidence of positive level IIb nodes within a range of 0 to 10.4%.12 Of note, there were only three episodes of isolated level IIb nodes out of 332 patients. None of the studies to date evaluate shoulder function when controlling for sublevel IIb dissection. In theory, dissection level IIb after level IIa dissection could lead to increased shoulder impairment postoperatively without improved oncologic control.
In summary, the literature supports the findings that sacrifice or dissection of key structures, particularly CN XI in level V, causes increased impairment.
12.3.4 Radiation and Chemotherapy
Adjuvant therapy for head and neck cancer also plays a role in head and neck cancer treatment-related morbidity. A multivariable analysis was performed using the NDII as the assessment tool to compare MRND sparing CN XI with SND, and adjuvant radiation or chemoradiation were independent predictors associated with shoulder impairment.8
Similarly, in a study of 25 who underwent SND levels I to III, and 86 who underwent extended SND (levels I–V) with sacrifice of SCM, the extended SND was associated with greater cervical range-of-motion deficit compared to SND levels I to III. In this study, 98 patients received radiation and the combination of extended SND and radiation produced additional impairments in cervical range of motion compared to extended SND alone. Radiation therapy as a single modality did not result in increased morbidity of cervical range of motion, mouth opening, swallowing, or lymphedema at 12 months posttreatment.13 Postoperative radiation therapy after neck dissection was associated with impairment of upper limb function and increase shoulder pain as measured by the NDII and Quick Disabilities of the Arm, Shoulder, and Hand (DASH) survey in a retrospective study examining 89 patients treated with SCM and CN XI–sparing neck dissections.14
Primary radiation or chemoradiation for treatment of oropharyngeal and nasopharyngeal carcinomas allows for examination of posttreatment effects. Examination of patients presenting with dysphagia greater than 5 years postradiation or chemoradiation for head and neck cancer revealed dysarthria or dysphonia, cranial neuropathies, and pneumonia in 76, 46, and 86% of patients, respectively.15 When looking at post–chemoradiation therapy (post-CRT) neck dissection for patients without complete response, nearly all patients had returned to a soft or regular diet by 2 years posttreatment with 10% of patients remaining gastrostomy tube dependent.16 The rate of gastrostomy tube dependence did not appear to be associated with posttreatment neck dissection
Overall, chemoradiation does not seem to cause shoulder morbidity as a primary therapy within 2 years of treatment. Use of chemoradiation in the adjuvant setting exacerbates the morbidity caused by neck dissection.
12.3.5 Sentinel Lymph Node Biopsy
Sentinel lymph node biopsy (SNB) is the standard staging tool for the management of melanoma, and it has been studied for use in early classification (T1–T2) of oral cavity squamous cell carcinoma demonstrating a negative predictive value of 95%.17 Few studies have been carried out examining the impairment of sentinel node biopsy versus elective neck dissection.
A retrospective comparison of 62 patients with early classification oral tongue lesions (T1–T2) who underwent either SNB (n = 33) or SNB followed by elective neck dissection (n = 29) demonstrated SNB was associated with less shoulder impairment than SNB followed by SND using the NDII and constant score as outcome measures.18
12.4 Assessment Tools for Neck Disability
There are many different tools available for the assessment of shoulder disability. 
Table 12.2 outlines the most commonly used questionnaires in the head and neck literature to assess shoulder function. Most of the assessment tools were designed to assess rotator cuff and/or glenohumeral joint disease.
The two patient-reported outcome (PRO) measures for assessment of shoulder function after neck dissection are the NDII and the DASH questionnaire. The NDII was specifically designed and validated in the head and neck cancer population.19 The DASH has undergone the most rigorous psychometric analysis and validation of any of the questionnaires listed in 
Table 12.2, and it is the best tool for comprehensive assessment of upper extremity function. A comprehensive review of the outcome measures has been conducted and the NDII was found to be the most appropriate assessment tool at this time.20
The DASH questionnaire was validated in the head and neck cancer patient population in a cross-sectional study comparing RND, MRND, and SND. Evaluation of the DASH by both physicians and patients met sensibility criteria, which means the DASH questionnaire asked questions appropriate for patients who underwent neck dissection. The DASH was also able to discriminate between the different types of neck dissection, and it was also validated in this patient population with high correlation of the DASH with the head and neck cancer patient-validated NDII.21 The Shoulder Disability Questionnaire (SDQ), NDII, and Shoulder Pain and Disability Index (SPADI) were also validated in a cohort of patients who underwent neck dissection; however, only the NDII was able to discriminate between types of neck dissection.22
In summary, there are two convenient questionnaires that are reliable constructs for evaluation of shoulder impairment following neck dissection—the NDII and DASH. Although the SPADI and SDQ are also validated in patients undergoing neck dissection, these measures do not appear to be as sensitive as the NDII and DASH.
