Neurofibromatosis type 1-associated plexiform neurofibromas of the neck: topography of lesions and surgical treatment data of 69 patients

The clinical data of NF1 patients with PNF of the neck were analyzed in a retrospective study. All patients were treated by the senior author for neck PNF between 2000 and 2019 at the Department of Oral and Craniomaxillofacial Surgery, Eppendorf University Hospital, Hamburg. All patients met the current diagnostic criteria of NF1 [12]. The surgical measures were primarily based on esthetic and functional reasons in order to reduce the size and shape of disfiguring tumors. Due to the size of the lesions, these were exclusively debulking procedures aimed at restoring the natural contours of the neck by contouring the region of interest. In these cases, esthetic and functional reasons were the predominant indications for surgical intervention. Due to the extent of the tumor, the interventions usually were partial tumor reduction. In some cases, rapid tumor growth was suspicious for malignancy of the mass (malignant peripheral nerve sheath tumor, MPNST) and justified surgical exploration. In the case of nodular tumors, in addition to the characterization of the main lesion, the assessment of the resection margins was the focus of the examination. However, “partial resection” also applies to nodular PNF with macroscopic complete removal, because it is not possible to determine the extent to which the remaining tumor nerve is affected by the nerve sheath tumor in the total length, which is inconspicuous on visual inspection and histological evaluation of border cuts, i.e. showing a benign nerve sheath tumor.

The data was determined from the medical files, operation reports, histological findings, medical reports, and digitally recorded in an anonymous form. The following findings were evaluated: sex, age, number and duration of surgical procedures of the neck, duration inpatient stay, surgery-related complications, histological findings, tumor localization, and number of tumor-affected regions per patient. The duration of the operations was divided into 10-min increments for a simplified overview of data.

In anatomical classification, the sternocleidomastoid and trapezius muscles divide the neck superficially into four regions (anterior, lateral, and posterior cervical regions and the sternocleidomastoid region), while deeper structures divide it into the so-called cervical triangles (trigonum submandibulare, submentale, caroticum, musculare, and omoclaviculare) [13, 14] (Fig. 1). Inferiorly, the neck is bounded by the clavicle and the superior border of the pars transversa of the trapezius muscle, and superiorly by the inferior border of the mandible, the mastoid process of the temporal bone, and the superior nuchal line of the occipital bone. The anatomical classification of the present evaluation considers the superficial 4 regions. The classification of the neck surface according to dermatomes differs from the anatomical classification mainly in the size of the sternocleidomastoid region and its demarcation from the anterior and posterior fields (Figs. 2 and 3). To avoid being influenced by the classifying lines of the skin divisions on the schematic drawings, the lines were inserted into the corresponding files only after the tumors had been drawn onto the digital head and neck surfaces. To include tumors that spread to the skull in the results, a field “adjacent head” was included in the evaluation in addition to the 8 localizations by dermatome/anatomical unit. In addition, photo documentation of the local findings of 39 patients from the archive of the senior author was evaluated. The visible extent of the tumor was manually transferred to a digitalized head and neck scheme [13, 14] (Fig. 1).

The plotted tumor extension was stored in Photoshop Version 2.5 (Adobe; Mountain View, CA, USA). Two sets of images, each modified for and adapted to our own investigations, each with four views (frontal, occipital, lateral right and left) were prepared for each patient. The two sets of figures show lines dividing the head into units or regions [13, 14]. The subdivisions of the skin surface in one set of images are based on the dermatomes of the examination area, the second on the defined anatomical units of the face and head. However, the initial recording of the tumor spread was made using a scheme without scaling in order not to influence the description of the tumor spread by pre-structured diagnostic fields. For each patient with existing photo documentation, the respective extent of the tumor was drawn manually on the scheme. This documentation resulted in different numbers with illustrations of the marked extent of the tumor. The accumulation of the number of tumor localizations was illustrated with a heat map. More commonly affected/treated regions are illustrated in red, less commonly in yellow, and rarely in green. However, the color coding cannot be firmly assigned to a specific number of tumors since different numbers of patients were considered for each side, that is, schematic neck projection. The respective pattern of colors only shows the relationship of tumor frequency for each side and cannot be related to the other sides regarding the number of tumors.

Extensive swelling of the surgical site and adjacent regions has occurred in almost all cases and is to be expected in a disease in which tumorous Schwann cells cause the accumulation of interstitial fluid by producing extravasation [15], a process that apparently can be greatly enhanced by mechanical irritation of the surgical intervention. As a result of tumor swelling during debulking procedures, the wound margins often swell, delaying epithelialization of the wound. These findings occur in almost all treatments and are not considered complications. Complications for the purposes of this study are primarily re-bleeding that required revisions, wide wound dehiscence due to unusual tissue swelling and suture insufficiency, and nerve damage.

All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. Prior to analysis, data were anonymized, and the investigators studying the radiographs were blinded to diagnosis and individual identity. The investigations of anonymized data were performed in accordance with Hamburgisches Gesundheitsdienstgesetz (Hamburg Health Services Act). This type of investigation does not require the approval of the local ethics committee. The examinations are part of a scientific thesis to fulfil the requirements for the doctorate at the faculty of medicine of the university (DML).

SPSS version 27 (IBM, Armonk, USA) was used to analyze the data. Absolute and relative frequencies, minimum, maximum, median, mean value (MV), and standard deviation (SD) were determined. The data was evaluated by calculating Spearman correlations because no normal distribution of the data could be proven.

Of the total 69 patients, 26 subjects (37.7%) were male, and 43 subjects (62.3%) were female. At the time of the first surgery, the youngest female patient was 3 years old, and the oldest was 63 years old. Females had a mean age of 29.59 years (SD = 14.38, median 30 years). The youngest male patient was 1 year old at the time of the first surgery, and the oldest was 75 years old. The mean age for males was 30.12 years (SD = 20.87, median 24.5 years). For men, the age at which surgery was performed is distributed over a wider range than for women. For men, the age range above the median is significantly larger than below. In a chronic disease of this localization, the need for surgical treatment in women may be concentrated in a shorter time interval of the life phase (Fig. 4).

Tumors were largely uniformly localized over the study area: Anatomocal regions (bilateral): sternocleidomastoid (32.5%), lateral cervical (25.0%), anterior cervical (25%), and posterior cervical region (17.5%). Anatomical units: left anterior cervical (14.4%), left sternocleidomastoid (13.5%), left lateral cervical (12.5%), right anterior cervical (12.5%), right sternocleidomastoid (11.5%), right lateral cervical (6.7%), and right posterior cervical region (5.8%). Co-involvement of other head regions in tumor infiltration was registered in 13.4%. None of the patients had impaired breathing due to the tumor. Two patients had severe scoliosis and reductions in lung volume resulting in instrumental respiratory support. Extensive tumors with tumorous infiltration and destruction of the musculature resulted in some cases in lowering of the shoulder of the affected side.

The number of regions affected per patient were: 1 (33.3%), 2 (18%), 3 (15.4%), 4 (25.6%), 6 (5.1%), 7 (2.6%).

In 69 patients, a total of 92 neck surgeries were performed for PNST. The number of surgeries performed on a patient ranged from one to four (MV = 1.33). Most patients, namely 51 patients (73.9%), were operated only once, while 14 patients (20.3%) were operated twice. The mean value for the difference in patients with multiple surgeries was 739.1 days. The minimum was 13 days, the maximum 4131 days.

In a total of 92 operations, no postoperative complications were recorded in 80 (87%). The most frequent complications were post-operative bleeding in three operations (3.3%), hematoma in two (2.2%), and severe swelling. The complications of fever, necrosis, sensory disturbance, wound secretion, and suspected thrombosis occurred once each (1.1%). In no case, a tracheotomy became necessary. The following diagram does not depict patients without complications to better illustrate relevant findings (Fig. 6).

On average, the inpatient stay was 6.91 days (shortest/longest stay: 1/32 day(s)), (Fig. 7).

The histological findings were reevaluated. In the histological findings obtained during the operations, 33 patients (32.7%) were found to have plexiform neurofibromas. The following diagram summarizes the tissue findings: Clearly, there is a difference in the description of tumors in terms of clinical and histologic terminology. The clinical assessment according to tumor extension and consistency (especially the bulky tumors) is often histologically assessed not as a plexiform tumor but as a diffuse neurofibroma. However, histologic evaluation of large specimen relies on selection of representative tissue samples (Fig. 8). Histologic distinctions did not affect the assessment of tumor biology, except for MPNST.

The photographic documentation of tumor outline of 39 patients constituted the heat map survey of the neck. Because tumors in some cases extend across multiple projections, patients are depicted as having a PNF on more than one body side of the schematic head drawings (front: n = 22, left: n = 24, right: n = 14, back: n = 23).

A visual comparison of the two skin surface classifications (anatomical and dermatomes) shows that the hotspots shown fit somewhat better into anatomical divisions, i.e., units. This is clearly visible in several examples: In the view from the right, both the hotspot below the ear and the second hotspot submandibular are divided by a white line according to dermatomes. In contrast, when classified by anatomy, both hotspots are outlined by the white lines. The same is true for the view from the left, in which the submandibular hotspot is divided by dermatomes, which is not the case in the division by anatomy. This phenomenon is also observed in the anterior and posterior views. Based on this, the manifestation of plexiform neurofibromas was analyzed in more detail for the individual surfaces of the anatomical units. Thus, it can be assumed that defining the anatomical neck PNF unit is relevant for surgical assessment of tumor spread. However, the tumors also cross the boundaries of the units in the anatomical classification of the skin surface (Figs. 2 and 3).

In a correlation analysis, the parameters of number of operations, number of affected regions, and length of inpatient stay after surgery were compared. None of the correlated parameters showed meaningful significance (Number of regions and number of surgical procedures: correlation coefficient (CC) = 0.032, p = 0.816; number of regions and inpatient stay: CC = 0.143, p = 0.302; number of surgical procedures and inpatient stay: CC = −0.106, p = 0.447).

No sex-specific clustering of tumor localizations was observed. Apparently, a few interventions are sufficient to achieve an improvement in appearance and function. Overall, the proportion of malignant tumors is low, which agrees with the distribution pattern of MPNST. The accumulation of surgical interventions to a certain age range corresponds to the known experience that tumor growth takes place particularly in youth and early adulthood and that the need for treatment already exists in this period of time. The boxplots indicate that the main phase of surgical treatment is preferably around 20–40-year-old patients. The correlation analysis shows no relationship between tumor extent (number of regions), number of surgical measures, and the duration of inpatient stay. Significant deviations from this estimate are to be expected when very large tumors that have grown far beyond anatomic units are treated surgically; when interventions are planned that are combined with skeletal/intraspinal measures of the region; and in cases with malignant peripheral nerve sheath tumors arising from neck PNF [16]. There are obvious differences in the tumor types to be treated in the respective specialist clinics. Thus, plastic surgical interventions are more common in some departments than in other departments that focus on the exploration of tumors of the brachial plexus. It is currently unclear whether the manifestations of the lesions differ depending on the genetic background of the patient. However, non-syndromic neoplasms were more common than NF1-associated ones in PNST of the brachial plexus [16]. Inpatient treatment time and complication pattern reflect follow-up characteristics of a disease with increased postoperative bleeding tendency and delayed wound healing [17, 18]. The length of the inpatient stay is explained by the delayed wound healing, the risk of secondary bleeding from the wounds, and the fact that outpatient specialist care for the patient at home is often not guaranteed. The complication rate is about 13% and justifies inpatient care of patients as well as individually adapted length of hospital stay. Bleeding from the PNF can influence the extent and duration of surgical measures [18, 19]. Wound healing results are esthetically satisfactory in many cases [20].

A graphical representation of tumor localization clustering illustrates approximately bilaterally symmetrical distribution of the tumors. Within the affected side of the body, patterns of tumor spread are recognizable, which were best described in this study by assignment to anatomical units. The examination confirms that the tumors have a segmental spread [10]. Furthermore, the preferential localization of the tumors in the sternocleidomastoid region is evident from the heat map. The clustering corresponds to data in the literature in which the surgical treatment results of brachial plexus PNF have been analyzed [15, 21, 22]. The description of cutaneous PNF distribution of the neck corresponds to results for the PNF distribution pattern developing in the extremities [23, 24]. In the regions of the trunk and extremities, segmental distribution of tumors across dermatomes and units was obvious [23‐25]. In contrast, description of tumor distribution according to dermatomes of the trigeminal nerve has proven useful for topographical classification of facial PNF [26, 27]. Whether the classification according to anatomical units for PNF of the neck is suitable for the analysis and treatment planning of NF1 patients in general should be tested on a larger study group. In addition to analyzing the frequency of tumor manifestations, the examination procedure is suitable for assessing the individual course of the disease as long as the tumors cause externally visible changes of the neck (e.g., to objectify tumor growth).

Histologic evaluation of the resected specimens confirmed the diagnosis of PNST in most cases [16, 17]. However, there are differences in the use of the term “plexiform” based on medical specialties [28‐34]. The histological distinction is based on the relationship between the axon and the tumorous Schwann cells. In histological definition, the diagnosis “plexiform” has no relationship to tumor size [24]. In contrast, in the clinical-diagnostic field, the term “plexiform” is predominantly used to describe both large tumors with infiltrative growth patterns (lumpy masses) and nodular PNSTs arising, for example, in the plexus (nodular PNF). The differences between the histologic definition of PNST and clinical practice indicate the current state of the diagnostic art, in which the term “PNF” is not clearly defined [8]. Among the variant PNSTs are especially sack-like tumors, often infiltrating only the subcutis, which are regularly classified histologically as diffuse neurofibroma/diffuse plexiform neurofibroma. Clinical assessment often leads to the designation of superficial neurofibroma [30, 33]. In most, but not all cases, this type of spread is considered a benign tumor [34]. The problem of selecting representative tissue samples for general diagnosis of a large tumor specimen has already been pointed out [23].

A recently published study mapped the distribution of plexiform neurofibroma on the body surface [25]. The number of head and neck PNF was high (19.2%). The PNSTs were addressed as diffuse plexiform neurofibromas. However, the diagnosis of PNF was not further specified by a histological or clinical definition of the entity. In addition, the tumor distribution was not classified according to dermatomes or anatomic units. Earlier studies suggest that head and neck PNF are very common tumor sites in NF1 patients requiring surgical intervention [3, 16, 17, 35, 36].

The imaging procedure was used to specify tumor localization and treatment needs of a group of patients suffering from specific tumors preferentially arising in NF1, a tumor predisposition syndrome. However, the selection criterion, i.e., analyzing surgically treated patients, must not lead to generalization of tumor distribution and its frequency in the region studied. However, the study may help to more precisely specify regions that are particularly stressful for patients and prompt them to seek surgical assistance and to support therapy planning.

The differences between histological evaluation of the tumor and its clinical assessment were revealed in this study (Fig. 8). Large, dewlap-like tumors that have been clinically assessed as PNFs on the basis of extension and disfigurement can be diagnosed as “diffuse” or “plexiform/diffuse” as well as “plexiform” neurofibromas on tissue analysis. These tumors frequently appear like superficial (often sagging) tumors due to the obvious deformation of the body region or shape. The radiologically defined distinguishing feature of superficial neurofibromas is that these do not invade the deeper layers of the body [29] and therefore constitute a separate entity [30]. However, the conclusion derived from this classification cannot be generalized. This type of tumor can invade the muscles of the neck. It is possible that this progressive phenotype of the superficial PNF is age-dependent.