CORR Insights®: What Is the Implant Survivorship and Functional Outcome After Total Humeral Replacement in Patients with Primary Bone Tumors?

Where Are We Now?

Orthopaedic management of malignant bone tumors of the extremities has matured considerably over the past few decades. Limb salvage following a radical oncologic resection, concepts previously considered incompatible, is considered as effective as amputation in terms of oncological outcomes like survival thanks to improved surgical techniques, modern chemotherapy and radiotherapy strategies, and advanced imaging.

Benefits of limb salvage over amputation include preserved distal extremity function, cosmesis, the mental benefit of having a restored limb, and the versatility of an endoprosthetic reconstruction [3]. However, replacement of the full length of a bone—as in total femoral or humeral replacement—brings on myriad challenges. The only option for total limb reconstruction is an endoprosthesis, since allograft reconstructions are contraindicated because of the lack of a junctional vascular supply needed to support unipolar osteoarticular allografts. But these large, metallic implants are at risk for infection and wound-healing complications, given that the soft tissue envelope has been traumatized by both cancer and surgery; the lack of organic muscular attachments around either joint makes mechanical complications more likely as well.

Despite all this, overall implant survival at 5 years following total humeral replacement is quite high. In one study of 20 total humeral reconstructions mostly for primary bone tumors, 19 had functioning, intact implants at a minimum of 2-years follow-up [4]. Another reported a similar 87.1% 5-year survival rate following total humeral replacement in 20 patients, but five developed an infection and nine were treated for dislocation or subluxation of their proximal humeral component [2].

In this month’s Clinical Orthopaedics and Related Research®, Schneider et al. [5] reported a 5-year implant survival rate of 74% in 31 patients with primary bone sarcomas treated with a total humeral replacement during a 19-year period. Many of the complications in this series were related to soft tissue healing and the extent of the muscular/capsular resection. The glenohumeral soft tissue envelope appears to carry particular importance, as the authors report a 5-year implant survival of 50% in patients who underwent an extraarticular resection compared with 89% following an intraarticular resection. While no flaps were used in these reconstructions, this finding may indicate that larger proximal resections merit more-aggressive muscular coverage and perhaps should be performed in conjunction with a plastic surgeon. The present study found that if the patient’s implant remained in service 1 year after surgery, 5-year postoperative survival was 95%. The authors also reported that those patients who survived 5 years after surgery had a mean Musculoskeletal Tumor Society (MSTS) score of 87% and a median American Shoulder and Elbow Surgeons Shoulder Score (ASES) of 83, representing good postoperative function. However, it should be noted that only 13 patients survived to this timepoint. This perhaps better than any finding highlights the paradoxical challenge of analyzing outcomes following total humeral replacement: The implant may be more likely to survive than the patient.

The findings reported herein have several clear clinical implications: (1) data can be used to guide not just global functional assessment postoperatively, but also what specific function is most critical to preserve quality of life; (2) patients should be forewarned about the potential for prosthetic complications within a year of surgery but encouraged that these risks go down markedly thereafter; and (3) extensive soft tissue resection around the shoulder girdle is associated with implant failure, and immediate soft tissue coverage should be considered liberally in patients where an extraarticular resection is oncologically necessary.

Where Do We Need to Go?

The encouraging implant-survival data presented by Schneider et al. [5] should change our focus regarding total humeral replacement from “can we?” to “should we, and in whom?”. Projected lifespan is of particular importance when considering limb salvage over amputation, both from a rehabilitation and a cost perspective. To this end, the authors of the present work chose only patients with primary tumors who underwent resection with curative intent. However, even in this population the survival at 5 years was < 50%. This should prompt us to reconsider the ideal timepoint for reassessing outcomes, as doing so 5 years after surgery fails to measure function in most patients (who may succumb to disease before that time). While it is unrealistic to expect much improvement at early timepoints, such as 6 to 12 weeks after the procedure, perhaps assessing patients 1 or 2 years after surgery would help us better to understand whether these procedures—which are associated with a substantial convalescent period and real surgical risk—are justified by measurable benefits to patients.

In addition to when function should best be assessed after a total humeral replacement, we also must question how it should best be evaluated. The findings of the present work [5] strongly suggest that shoulder function does not have a substantial impact on patient-reported functional outcomes following total humeral replacement; we know this because a particular strength of Schneider et al.’s work is their meticulous documentation of what shoulder anatomy and innervation was sacrificed during their oncologic resections. The deltoid muscle was sacrificed in 42% of patients, the radial nerve in 16% of patients, and the axillary nerve in 35% of patients. The same degree of shoulder dysfunction was seen by Kotwal et al [2], who reported that none of the patients at max follow-up in their set could actively elevate their arm against gravity to the level of the patient’s face. Tsuda et al. [6] found comparable MSTS functional scores in patients treated with endoprosthetic reconstruction as with resection arthroplasty of the shoulder following extra-articular tumor resection, further supporting the idea that the shoulder is not a main driver of functional outcomes in these patients. It is unclear why the shoulder is generally inconsequential to patient-reported function following tumor resections around the girdle. This could be due to the overwhelming importance of hand, wrist, and elbow salvage; extensive patient preparation to the likelihood of permanent shoulder dysfunction; or the poor function of current endoprosthetic reconstructions. This has not been a focus of prior studies and may merit further investigation.

With shoulder function in mind, the reverse total shoulder modular component used in a small subset of patients in the present work [5] may be a gamechanger for restoring shoulder function, and it may allow patients to expect measurable restoration of abduction postoperatively. Unfortunately, the use of a reverse total shoulder construct on only a small subset of patients (three) precluded objective analysis in this work. However, regardless of the potential for shoulder function restoration, the good or excellent MSTS scores reported for all surviving patients by Schneider et al. suggest that improving shoulder function should be secondary to preserving elbow and wrist function. In light of this, I agree with the authors that there may be a role for early involvement of upper extremity specialists during shoulder resection and reconstruction, who may be able to improve the patient's ultimate function by means of well-planned tendon transfers and restorative procedures based on intraoperative findings. We may also choose to analyze total humerus outcomes with more hand/wrist-focused patient-reported outcome metrics, such as the Disabilities of the Arm, Shoulder and Hand questionnaire or the Michigan Hand Questionnaire [1].

How Do We Get There?

Megaprosthesis research is challenging. The sample sizes are small, the technologies are rapidly changing and often involve custom patient-specific implants, the metrics used to evaluate outcomes are heterogeneous, and individual studies are mostly retrospective. Further complicating high-quality research involving patients with cancer is the fact that patient-reported outcomes are influenced by disease progression, and therefore interpreting 5-year outcome data in a population that only had a 44% survival rate to that point (such as in the present work [5]) may not add much information about patient function post-healing but predecline. Total humeral prostheses are uncommon, and few are performed yearly. It is therefore very difficult to perform a nuanced analysis in any single study. We are still at an early phase of understanding the indications and target outcomes for total humerus reconstructions, as illustrated by Schneider et al.’s primary question focusing on implant survival rather than quality of life and function.

The authors of the present study provide one of the largest cohorts of total humeral replacements yet published, and they include functional scores within their analysis. That this group is the largest that I know of (and only includes 33 patients across two paradigmatic reconstructive shifts over a 19-year study period) helps us understand the difficulty in studying these reconstructions. While the current body of evidence demonstrates the feasibility and likelihood of a total humeral replacement surviving after management of a primary bone tumor, it is ultimately unable to provide feedback on patient selection and prognostic factors using the currently available sample.

Unfortunately, opining that further research, larger samples, more uniform measurements, and consistent technology can improve our understanding of total humeral replacements likely illustrates a case of wishful thinking. The shear rarity of this procedure likely precludes a compendium analysis, and the heterogeneity of a multi-institutional work would dilute its quality. We must, therefore, be selective about our way forward and ask questions capable of study within a limited sample of patients. Of particular interest is determining the utility of a reverse total shoulder component in a total humeral replacement construct, which in turn might permit the assessment of shoulder function. We could then come to some conclusions about whether it truly has only a marginal impact on patient-reported outcome. An assessment of soft tissue coverage strategy may improve implant survival after extraarticular resection, which is a key weakness. Finally, a comparative reassessment of functional outcome studies at short (1-3 months) follow-up in this unique population may help us understand what aspects of life can be “as normal” for patients at a time period during which survival and function is most important and which areas should be the focus of rehabilitation and future innovation.