Differential Diagnoses of Pediatric Upper Limb Masses

We performed a retrospective review of all patients aged ≤18 years presenting with an upper limb mass to our institution over a 5-year period. Our hospital is 1 of the 2 dedicated children’s hospitals in the country, with 3 fellowship-trained hand surgeons who manage pediatric upper-extremity tumors. We included all children who presented with a clinically palpable or radiologically evident mass. Although skin neoplasms behave differently from true mesenchymal tumors, we included skin and nail lesions for 2 reasons. First, several patients elected to see a hand surgeon instead of a dermatologist or pediatrician, primarily because of the anatomic site of involvement. This is a more realistic reflection of the broad spectrum of cases that may be encountered in certain hand surgery practices. Second, pigmented nail lesions may be indicative of a subungual melanoma, although the most common cause in children is a melanocytic nevus. Such cases are occasionally referred to hand surgeons because of the need for a nail matrix biopsy to establish a diagnosis. We reviewed the electronic medical records and recorded information on patient demographics, clinical features, radiological investigations, intraoperative findings, and histological features. This study was approved by the institutional review board of National University Hospital.

We identified 140 pediatric patients with 141 upper limb masses. There were 63 (45%) males and 77 (55%) females (Table 1). The mean age at presentation was 12 years (SD, 5 years). Seventy-five (53%) tumors involved the right upper limb. The dominant hand was involved in 61 (43%) cases, the nondominant hand was involved in 50 (36%) cases, and the data were not available in 30 (21%) cases. The tumors were localized to the arm in 14 (10%) cases, the forearm in 12 (9%), the wrist in 63 (45%), the hand in 15 (11%), and the digits in 37 (26%; Fig. 1A). The middle finger and thumb were the most commonly involved digits (Fig. 1B). There was 1 patient with masses in both the thumb and small finger.

The clinical presentation varied, and some patients had more than 1 presenting complaint (Table 2). Seventy-four patients presented with painless masses, and 63 had painful masses. Joint stiffness was present in 6 patients. Two patients reported mechanical features; 1 patient experienced clicking on flexion of her middle finger and another had localized pruritus. Neurological features such as numbness, paresthesia, and weakness were present in 7 patients. Vascular features were seen in 8 patients, and these included discoloration, cold intolerance, and reduction in size on the elevation of the limb. Three patients had masses that were discovered incidentally when they were being evaluated for unrelated issues. None of the patients in our series presented with pathological fractures.

There were 128 soft tissue tumors and 13 tumors arising from bone (Fig. 2). The most common tumor was a ganglion cyst, comprising almost 70% of the soft tissue tumors (Table 3). Fifty-eight (68%) patients diagnosed with ganglions underwent additional imaging. Among them, 29 had ultrasonography (USG), 5 had MRI scans, and 2 had both an MRI scan and USG. Only 10 (12%) of these patients underwent surgery, and the main indications were localized discomfort, enlargement, and cosmetic concerns. Open marginal excision was performed in 9 children, and 1 patient had arthroscopic ganglionectomy. A histological analysis was performed in 7 of these patients, and all of these analyses confirmed the clinical diagnosis.

The remaining 43 soft tissue tumors were mostly benign neoplasms (Table 3). However, open surgery was required in 63% of cases. The most common indications for surgery in this group were localized symptoms, evidence of clinical progression (eg, enlargement), and uncertainty regarding the diagnosis. These goals were not mutually exclusive, and often surgery fulfilled several of these objectives. A marginal excision was performed for symptomatic masses that were well localized and for which the clinical and radiological investigations strongly suggested a benign diagnosis. A critical point here is that the surgeon must be prepared to take an incisional biopsy and not proceed with further soft tissue dissection if atypical features are encountered during surgery, such as tumor heterogeneity, exophytic areas, and excessive vascularity. In such cases, it is safer to obtain a histopathological diagnosis before proceeding with definitive resection. In our series, 20 patients had marginal excisions, and these were all benign tumors. These masses were typically small, ranging from 2 to 3 cm in the hand and less than 2 cm in the digits. Open intralesional procedures were performed when the goal of surgery was to obtain a diagnosis and concomitantly reduce the size of a benign mass to alleviate local symptoms, such as for chondroma and junctional melanocytic nevus. In children where there was doubt regarding the diagnosis and the clinical assessment and radiological investigations suggested a locally aggressive or malignant process, we took a staged approach, with an initial biopsy followed by a definitive resection. A percutaneous core biopsy was preferred, and image guidance using USG or a computed tomography scan helped to sample the representative areas of the tumor. The 3 children in our series with a locally aggressive or malignant diagnosis had biopsies before definitive surgery: 2 underwent wide excision and 1 had a radical excision. A wide excision involved complete resection of the tumor, together with the surrounding reactive zone. A radical excision was one in which the entire soft tissue compartment was removed.^, One of these patients also required additional reconstructive surgery. This patient had a high-flow arteriovenous malformation involving most of the extensor compartment and required tendon transfers to restore wrist and finger extension after tumor resection. The specific diagnoses and biological behaviors of these tumors are shown in Table 4. Of these 43 soft tissue tumors, 39 (91%) were benign, 3 (7%) were malignant, and 1 (2%) was intermediate in nature (Table 3). “Intermediate” tumors display an infiltrative and a locally destructive growth pattern. They may recur locally but do not metastasize. In our series, a child presented with a painless, solid mass that proved to be an angiomatoid fibrous histiocytoma, a rare soft tissue tumor of intermediate biological potential. The final histopathological diagnosis in approximately half the solid tumors that were excised did not correlate with the initial clinical diagnosis, but only 1 eventually proved to be a malignant neoplasm (Fig. 3).

Most (10 of 13) of the bone tumors seen were benign neoplasms (Table 3). There were 3 malignant tumors: 2 osteosarcomas and 1 metastatic tumor from a primary clear cell carcinoma of the kidney (Tables 5, 6). Surgery was performed in 6 (46%) patients with bone tumors. The principles of surgery for bone tumors were similar to those for solid soft tissue masses. Intralesional procedures were performed to obtain a definitive diagnosis or to remove a symptomatic intramedullary tumor. In our series, 2 patients had open intralesional procedures, including an incisional biopsy for a metastatic clear cell carcinoma and debridement, arthrodesis, and bone grafting for a symptomatic carpal boss. Marginal excisions were performed for well-circumscribed benign masses such as osteochondroma and surface lesions. Wide and radical resections were performed for biopsy-proven malignant tumors. There was 1 above-elbow amputation performed for an osteosarcoma of the forearm. One patient required reconstruction of the proximal humerus with a nonvascularized fibular autograft after the resection of a large osteochondroma. All histological analyses confirmed the initial clinical diagnosis.

The details of the 6 patients with malignant tumors are summarized in Table 6. Five patients had primary malignancies, and 1 patient presented with a hand mass that was a metastatic deposit from a renal clear cell carcinoma. Two of these patients returned to their native countries for further medical treatment.

Upper limb masses are commonly encountered by hand surgeons. Despite this, there is a scarcity of epidemiologic studies on upper limb tumors, particularly in the pediatric and adolescent population. There have been several small studies describing the differential diagnoses of pediatric upper limb masses and 1 large study that only reviewed children with hand and wrist tumors that were excised.^{,  ,}  We present a descriptive study that highlights the diagnoses of pediatric upper limb tumors in all children who were referred to our care either for a palpable mass or a tumor that was radiologically evident.

Our study illustrates several important points regarding the differential diagnoses and evaluation of pediatric upper-extremity tumors. First, most of the tumors in our series were benign, and locally aggressive and malignant neoplasms comprised only 9% of all cases. These findings are consistent with those of previous reports.^, Second, ganglions were the most common masses seen in our study, accounting for 60% of all cases and 66% of soft tissue tumors. This is also similar to the adult population, where ganglions account for 50% to 70% of the upper limb tumors.^,, Most of the ganglion cysts were accurately diagnosed by clinical examination and USG, and only 7 patients required an MRI to evaluate ganglions that were deep, in unusual locations, or multiloculated. However, there are considerable differences in the clinical presentation of these cysts across different series. Half of our patients with ganglion cysts had pain. This finding is similar to that of a study by Tang et al, in which almost 43% of adults had symptomatic ganglions but differs from the findings of studies by Westbrook et al and Barnes et al, in which symptoms were observed in 20% to 26% of the patients. Nevertheless, only 12% of the ganglions in our series eventually underwent surgery because of persistent pain, mechanical effects (eg, discomfort with extreme wrist extension), and cosmetic concerns, which is similar to the previously published figures.^, Although we did not have long-term prospective data, it is likely that most children with symptomatic ganglions experience an improvement or complete resolution of pain and local mechanical symptoms, a finding that has been well documented in adults. A recent study reported that only a small percentage of children with persistence of symptoms eventually undergo surgery.

The third point is that there is a long list of differential diagnoses for nonganglion soft tissue tumors, many of which are not truly neoplastic processes. Schultz et al reported that trauma, infection, inflammatory diseases, degenerative processes, and congenital abnormalities may present as upper limb masses that mimic tumor. Similar to the report by Colon and Upton, vascular malformations and pigmented villonodular synovitis were the other relatively more common diagnoses in our series, and malignancies were rare. However, we noted that more than half of patients with solid soft tissue masses who eventually underwent surgery had a histological diagnosis that differed from the original clinical impression. This may seem alarming, and we examined this issue in greater detail. In these cases, the clinical and radiological features were not specific, and it was impossible to determine the accurate diagnosis based on clinical and radiological assessments. This frequently occurred in tumors of a vascular nature, such as angioma and arteriovenous malformation, where the attending physician could correctly label the mass as a vascular lesion, but no further subcategorization was possible. Another scenario is in the case of small (<5 mm), solid masses, where the radiological features are not diagnostic, such as neurofibroma, dermatofibroma, desmoid tumor, and fibroma. Nevertheless, the most important point is that the attending physician was able to make the correct assessment regarding the behavior of the tumor in all but 1 of the 27 cases where surgery was performed. In the patient where malignancy was not initially detected, there was a forearm mass that did not present with classic “red flag” features such as pain or rapid enlargement, and the initial impression was of a benign vascular tumor. A biopsy was recommended to obtain a definitive diagnosis, but the parents initially decided against any invasive interventions. The child subsequently presented 1 year later with lymphadenopathy, and further evaluation revealed rhabdomyosarcoma with advanced systemic disease (Table 6). This case illustrates the challenges of making a diagnosis in children and adolescents. We noted that children aged <12 years were usually not able to comply with the stringent conditions necessary for MRI, and most of them required GA. This process is cumbersome and resource-intensive because a fully functional pediatric anesthesiology team must be set up in the radiology suite. Furthermore, young children do not tolerate minimally invasive procedures, such as core biopsy, under local anesthesia. This is inevitably a barrier for parents and surgeons, and may wrongly cause them to observe these tumors. Finally, a higher proportion of children and adolescents with nonganglion soft tissue tumors and bone tumors in our series required surgery than those with ganglions, even though the prevalence of these conditions was lower than that of ganglion cysts. The main indications for surgery were persistent local symptoms, evidence of progressive enlargement, and uncertainty regarding the diagnosis.

We have used our observations to develop a clinical algorithm that can help physicians make an accurate diagnosis with judicious use of investigations and biopsy (Fig. 4). A thorough history eliciting the symptoms, duration, progression, and antecedent causes, such as trauma, is essential. The physical examination must aim to determine the nature of the mass and the plane of involvement using simple maneuvers such as manual palpation, joint motion, and active muscle contraction. We routinely attempt to transilluminate masses using a pocket flashlight and examine the draining lymph nodes in every child. If the clinical features are insufficient to differentiate between a soft tissue or bony mass, we obtain a set of plain radiographs. Our current practice relies heavily on USG to better evaluate soft tissue masses because it reliably distinguishes between solid, vascular, and cystic masses; is affordable; and is well tolerated even by young children. We obtain MRI scans for solid and vascular soft tissue masses and for bone tumors that suggest a moderately or rapidly evolving process based on well-described radiographic characteristics. We obtain these scans under GA in children under 12 years, in older children who prefer GA, and in individuals with specific issues such as claustrophobia. We perform a biopsy for any mass that appears suspicious on MRI. Features that suggest a locally aggressive or malignant process include a heterogeneous appearance, contrast enhancement, surrounding edema, and loss of intervening tissue planes. We prefer to observe asymptomatic masses that do not show suspicious features and recommend the excision of ganglion cysts and other benign soft tissue masses that are associated with symptoms such as pain or soft tissue impingement. For bone tumors, an additional consideration is the risk of fracture. Lesions with a higher risk of fracture—based on features such as pain, focal tenderness, and considerable cortical involvement on plain radiographs—should be managed appropriately with a period of immobilization or internal fixation. Finally, we have found it extremely helpful to have access to a multidisciplinary tumor board comprising musculoskeletal oncologic surgeons, radiologists, oncologists, and pathologists to discuss complex cases.

The main limitations of this study were its retrospective nature and the inherent issues such as incomplete data and selection bias. Although our center is 1 of the 2 national children’s hospitals, we also receive patients from abroad. In addition, 1 of the hand surgeons in our practice is dual-certified in musculoskeletal oncology, and the definitive treatment of the more complex cases was performed by the same surgeon. Thus, our findings may not be entirely applicable to a different practice setting. Finally, Colon and Upton reviewed 349 pediatric hand tumors over an 18-year period. Our study was conducted over a shorter period, and we reviewed a smaller number of cases. Nevertheless, we also included patients who were managed nonsurgically and believe that the distribution of diagnoses is more representative of the range of conditions that might be expected in a tertiary referral center, notwithstanding the selection biases that may be present.