1 Introduction

Giant cell tumours (GCTs) of the bone are benign, locally aggressive skeletal neoplasms which were first described by Travers and Cooper in 1818.¹ They constitute approximately 5% of all primary and 20% of all benign bone tumours, although higher numbers (20%) have been reported from South-east Asian countries.^2–4 GCTs most commonly occur in the age group of 20–40 years, with a slight female predisposition.^2,5 Classically, they appear as osteolytic, eccentric, expansile lesions involving the metaphysioepiphyseal region of long bones on radiograph. The distal end radius (DER) is the third most commonly reported location, after distal femur and proximal tibia and it accounts for nearly 10–15% of the total burden of disease.^6–9

GCTs have around 0.16–4% risk of pulmonary metastases.^10,11 Lesions arising from the DER have much higher incidence of pulmonary metastases than other locations, accounting for approximately 26% of all metastasizing GCTs which involve the lung.¹² The most common explanation for this unusual behaviour is the increased risk of local recurrence (LR) associated with DER.¹⁰ GCTs have an overall recurrence rate of around 12–50%^13,14 based on the modality of treatment – 27–65% following simple curettage; 12–27% following extended curettage with adjuvants and 0–12% following wide resection.^15,16 Moreover, the risk of LR is up to 50% at the DER as compared to 28% at the proximal tibia and 13% at the distal femur, respectively.⁸ The need to mitigate the risk of LR while at the same time minimising functional imitations following surgery, are the main challenges associated with surgical management of DER GCTs. While wide resection has been shown to have a lower recurrence rate as compared to curettage, there is still a great deal of inconsistency in literature with regards to the overall functional outcome following either procedure.^13,14,17 Additionally, there is also paucity of evidence with regards to the most optimal reconstructive option following wide resection. While wrist arthroplasty (AP) allows motion at the wrist, arthrodesis (AD) has been shown to preserve grip strength (GS) and improve joint stability.^9,18,19

Therefore, the purpose of this study was twofold- (1) To make an objective comparison between intralesional curettage (IC) and wide excision (WE) as definitive treatment modalities of DER GCTs, using well defined oncological and functional scores, and (2) To evaluate the two types of reconstruction following WE and determine which of them have a better outcome.

2 Materials and methods

No ethical clearance was needed for this study as it was a systematic review of literature.

3 Results

3.1

3.1 Screening process

A total of 69 records were identified in the initial phase of screening. These included 67 studies from the PubMed database and one single study from the Cochrane Library. Additionally, one more study was added after a thorough scan of the references of the primary search results.

Seventeen articles were brought forward into the second phase of the study selection process, whose full-length texts were then closely scrutinized. Finally, eleven studies satisfying our inclusion/exclusion criteria were included in the review.

A similar search strategy was adopted for the second part of our study viz. wrist AP vs AD. After final screening, four articles were selected for this part of the study.

Two studies were common to both pools.^7,9

The PRISMA flowchart depiction of the literature search is presented in Figs. 1 and 2.

Close

Fig. 1 Flowchart depicting screening and study selection (curettage vs wide excision).

Export to PPT

Close

Fig. 2 Flowchart depicting screening and study selection (arthroplasty vs arthrodesis).

Export to PPT

3.4

3.4 Recurrence rates

A total of 42 LRs were reported amongst the 145 patients undergoing curettage (29%) as against four in the resection group (2.9%) (Table 3)

Table 3 Oncological and functional outcome parameters following curettage and wide excision.

SNO.	STUDY	INTRALESIONAL CURETTAGE (IC)								WIDE EXCISION (WE)
		SS	LR	ROM_FE	ROM_ UDRD	ROM_ PS	GS (%)	DASH	MSTS	SS	LR	ROM_FE	ROM_ UDRD	ROM_ PS	GS (%)	DASH	MSTS
1.	VanderGreind7 (1993)	5	0	–	–	–	–	–	–	17	0	–	–	–	–	–	–
2.	Sheth17 (1995)	18	5	–	–	–	66 ± 25	–	–	8	0	–	–	–	57.5 ± 20.1	–	–
3.	Cheng25 (2001)	6	0	–	–	–	89 ± 6.2	–	–	6	0	–	–	–	70 ± 6.5	–	–
4.	Harness5 (2004)	31	13	–	–	–	–	–	–	15	2	–	–	–	–	–	–
5.	Panchwagh26 (2007)	12	-a	–	–	–	–	–	82.4	12	-a	–	–	–	–	–	73.7
6.	Kang27 (2010)	9	2	126	–	–	83	9	–	6	0	–	–	–	58	21	–
7.	Chachairujira28 (2011)	2	1	–	–	–	–	–	–	8	0	–	–	–	–	–	–
8.	Wysocki9 (2015)	20	6	93.2 ± 20	25 ± 8.8	–	97.6	7	90	19	0	107.5 ± 19.1	42.5 ± 7.9	–	76	6.4	89.4
9.	Zhang15 (2016)	2	1	–	–	–	–	–	–	18	1	–	–	–	–	–	–
10.	Mozaffarian29 (2018)	6	4	138 ± 8.1	44 ± 3.8	166	87 ± 8.1	–	–	7	0	115 ± 8.6	34 ± 2	152	69 ± 5.8	–	–
11.	Abuhejleh10 (2019)	34	10	–	–	–	–	–	95.7	23	1	–	–	–	–	–	77.1
TOTAL		145	42							139	4

a Overall recurrence rate was 32%.

Quantitative analysis revealed the difference to be statistically significant (OR 8.6 [95% CI, 3.4, 21.75]), in favour of WE (Fig. 3).

Close

Fig. 3 Forest plot depicting the comparison in recurrence rates (RR) between the procedures.

Export to PPT

No recurrence was detected in either of the two groups (AP and AD) undergoing excision.

3.5

3.5 Functional outcome

(1)ROM:

The average ROM in flexion/extension (ROM_FE) was 119.1 ± 23.2° (IC) as compared to 111.25 ± 5.3° (WE). Likewise, for ulnar and radial deviation (ROM_UDRD), the values stood at 34.5 ± 13.4°, in the curettage group, and 38.25 ± 6° in the wide resection group. Quantitative analysis of these studies failed to detect any significant difference {(SMD 4.59 [95% CI, −31.96,41.14]) and (SMD -3.67[95% CI, −30.62, 23.28])} in the overall ROM following either of the two techniques (Fig. 4).

Close

Fig. 4 a, b Forest showing the comparison of range of motion (ROM) in the flexion-extension and radial-ulnar deviation arcs between curettage and wide excision.

Export to PPT

Out of the four studies directly comparing wrist AP with AD, only two^9,19 had made use of ROM in the pronation/supination arc as a functional parameter. These values have been summarized in Table 4 Data analysis of these parameters did not favour any one mode of reconstruction over the other (SMD -8.71 [95% CI, −49.97, 32.54]) (Fig. 5).(2)GS:

Table 4 Oncological and functional outcome parameters following arthroplasty and arthrodesis.

SNO.	STUDY	ARTHROPLASTY (AP)							ARTHRODESIS (AD)
		SS	LR	ROM_FE	ROM_PS	GS (%)	DASH	MSTS	SS	LR	ROM_FE	ROM_PS	GS (%)	DASH	MSTS
1.	VanderGreind7 (1993)	3	0	–	–	–	–	–	14	0	–	–	–	–	–
2.	Zhu19 (2013)	7	0	71.6	140 + 14.7	59.2 + 13.6	–	86.3	7	0	55.9	127.6 + 14.2	76.5 + 4.6	–	85.3
3.	Wysocki9 (2015)	4	0	107.5	127.5 + 12.7	69.4	20	87	15	0	–	157.2 + 14.5	77.6	3	90
4.	Qu18 (2018)	13	0	–	–	40	17	83	8	0	–	–	71	8	93
TOTAL		27	0						44	0

Close

Fig. 5 Forest showing the comparison of range of motion (ROM) in the pronation-supination arcs between arthroplasty and arthrodesis.

Export to PPT

Average GS in the IC group was 84.5 ± 11.65 (range, 66–97.6) % while for WE the value observed was 66.1 ± 8.1 (range, 57.5–76) %. Quantitative analysis revealed a significantly more optimal GS in the curettage group (SMD 18.08[95% CI, 13.78, 22.37]) (Fig. 6).

Close

Fig. 6 Forest depicting the comparison of grip strength (GS) values between curettage and wide excision.

Export to PPT

Likewise, the results of comparison between the modes of reconstruction showed a mean GS of 56.2 ± 14.9 (range, 40–69.4) % for AP and 75 ± 3.5 (range, 71–77.6) % for AD (Table 4).(3)DASH and MSTS scores:

A higher (poorer) average DASH score was reported in the WE group (13.7 ± 10.3) as compared to IC (8 ± 1.4). The same for AP vs AD showed better results in the latter.

Analysis of MSTS scores depict a trend more in favour of intralesional excision [89.4 ± 6.7, range 82.4–95.7 with IC vs 80.1 ± 8.3, range 73.7–89.4 with WE) and wrist fusion (89.4 ± 3.9, range 85.3–93 with AD vs 85.4 ± 2.1, range 83–87 with AP).

Values have been summarized in Tables 3 and 4, respectively.

3.6

3.6 Complications

The two most common complications in the IC/WE group were – (i) Fracture with/without associated non-union of the graft-host bone junction, and (ii) Degenerative arthritis of the wrist. The incidence of nerve injury was reportedly higher in patients undergoing curettage (Table 5).

Table 5 Complications following curettage and wide excision.

	INTRALESIONAL CURETTAGE (IC)	WIDE EXCISION (WE)
STUDIES	•VanderGreind 19937
	•Sheth 199517
	•Cheng 200125
	•Harness 20045
	•Panchwagh 200726
	•Kang 201027
	•Chachairujira 201128
	•Wysocki 20159
	•Zhang 201615
	•Mozaffarian 201829
	•Abuhejleh 201910
COMPLICATIONS
Fracture ± Non union	7	28
Carpal subluxation	3	4
Neurovascular injury	4a	2b
Wound necrosis and infection	2	6
Metastases	2	2
Wrist arthritis	8	12
Hardware related complaints	1	3
Miscellaneousc	2	1

a Two superficial radial and two median.

b One RSD and one transient CPN palsy.

c Two cases requiring tendon transfer in the IC group and one instance of pneumonia in the WE group.

In the second group, the incidence of graft fracture and non-union was higher in patients of AD. On the other hand, carpal subluxation and arthritis was seen more frequently following AP (Table 6).

Table 6 Complications following arthroplasty and arthrodesis.

	ARTHROPLASTY(AP)	ARTHRODESIS (AD)
STUDIES	•VanderGreind 19937
	•Zhu 201319
	•Wysocki 20159
	•Qu 201818
COMPLICATIONS
Fracture ± Non union	3	10
Carpal subluxation	8	0
Neurovascular injury	1a	1b
Wound necrosis and infection	1	0
Wrist arthritis	6	0
Miscellaneousc	0	1

a Transient CPN palsy.

b c/o RSD.

c Pulmonary metastases in the AD group.

Quantitative analysis of the complication rates showed a more favourable outcome with curettage vis-à-vis resection (OR 0.3[ 95% CI, 0.14, 0.62]) and AD vis-à-vis AP (OR 6.36[ 95% CI, 1.72, 23.52]) (Figs. 7 and 8).

Close

Fig. 7 Forest plot depicting the comparison in complication rates between curettage and wide excision.

Export to PPT

Close

Fig. 8 Forest plot depicting the comparison in complication rates between arthroplasty and arthrodesis.

Export to PPT

4 Discussion

GCTs are one of the most common musculoskeletal neoplasms seen in current orthopaedic practice. Although considered benign, these tumours are locally aggressive with a small propensity for pulmonary metastases.^2,13,14,29 Moreover, it has also been seen that lesions located in the DER have a higher risk of LR than those at other sites.^9,10,15 Probable contributing factors for the above include anatomic geometry of the DER, paucity of surrounding muscle cover, complexity of the distal radio-ulnar joint and close proximity to the neurovascular bundle and flexor and extensor tendons of the hand.⁹ These lesions pose significant treatment challenges for the operating surgeon in obtaining a balance between an oncological clear margin and best possible function. Past studies [ 7,9,10,13,14,26] have attempted to shed light on this dilemma and prescribe clear cut guidelines towards the surgical management of DER GCTs. They have shown that while curettage preserves wrist joint function and is a less morbid procedure, wide resection has a lower rate of recurrence, especially in Campanacci grade III tumours.^13,14 However, there has been a substantial addition to the literature^9,10,15,29 since Pazionis et al.¹³ published their findings in 2013 which we believe warrants a review.

Our study demonstrated a significantly lower recurrence rate with WE (2.9%) as compared to IC (29%). Similar findings were reported by Pazionis et al.¹³ (IC- 31% and WE-8%) and Liu et al.¹⁴ (IC- 31% and WE -7%). A subgroup analysis for Campanacci grade III lesions by the latter revealed the risk of LR with WE to be 3% vis-à-vis 28% for curettage alone. However, presence of a grade III lesion should not be taken as a reason to exclude treatment by curettage. According to Cheng et al.,²⁶ the first line of surgical management in such cases was still intralesional excision or curettage, especially in patients where the tumour did not invade the wrist joint, destroy more than 50% of the cortex or break through the cortex with an extraosseous mass in more than one plane. This was further reinforced by Kang et al.²⁸ who described Campanacci grade III(p) as a subgroup of grade III tumours wherein there was a single site of tumour containment by the pronator quadratus at the pre-operatively identified location of palmar cortical perforation. However, this narrative was countered by Mozaffarian et al.³⁰ They disagreed with Cheng et al.‘s assessment of performing an intralesional excision in all patients with less than 50% cortical destruction, owing to the relatively high recurrence rates observed with curettage by them in their own study (66.7%). The currently accepted figure for LR following IC is 31–35% and for WE 0–8%, respectively.^14,31 While some studies^26,28 do recommend intralesional excision for select grade III lesions, our consensus was to reserve curettage for only grade I and II GCTs of the DER. One reason was that the latter tend to have a more delayed clinical presentation than similar lesions at other sites, for instance, weight bearing long bones, and hence the need for a more aggressive line of management.¹⁰ Another reason was the functional compensation provided by the uninvolved ipsilateral upper limb which compensated to some extent for the morbidity associated with resection.¹⁰

A retrospective analysis of fifteen grade III GCTS of the DER²⁸ revealed significantly higher GS and improved Visual Analogue scale (VAS) scores in patients undergoing intralesional excision as compared to those who were treated with resection. Similarly, Cheng et al.²⁶ compared six patients treated with curettage with another six managed with wide resection and osteoarticular allograft. At a mean follow up of six years, it was found that patients belonging to the second group retained 69% of wrist ROM and 70% of the original GS. In stark contrast, no measurable differences were seen in the curettage group. However, Sheth et al.¹⁷ were able to demonstrate equivalent functional results while comparing the two techniques. Likewise, a prospective evaluation of thirteen DER GCTS by Mozaffarian et al.³⁰ showed patients undergoing wide resection retaining an acceptable ROM in the flexion/extension and pronation/supination arcs which was 83% and 92% of the curettage group. The authors attributed it to an intense but supervised rehabilitation program along with their efforts in maintaining ulnar variance at all costs. Functional results obtained by Wysocki et al.⁹ showed identical MSTS scores between the intralesional excision and resection/AD cohorts. Quantitative data analyses from our review failed to detect any significant difference between either groups, although, higher GS was seen with curettage. However, due to the variability in the outcome assessment measures used, we were unable to reach a definite conclusion regarding functional comparison between curettage and wide resection and believe large, well-designed, multicentric trials are needed to better determine the functional difference between both procedures.

Complication rates associated with distal radius resection vary from 29 to 100%.^5,7,17,26,28 These include non-union/malunion of the graft-radius junction (12–38%),^7,17,32 graft fracture (13–29%),^7,17 carpal subluxation (12–67%)^7,26 and degenerative arthritis of the wrist joint (13–50%).^5,26 In our review, the overall rates observed were 41.7% with WE and 20% with IC, respectively. At the same time, it was also noted that intralesional excision had a higher incidence of neural injury than resection and AD. This included two cases of median and one case of superficial radial nerve injury,¹⁷ besides one incident of superficial radial nerve irritation.²⁸ These findings, in all likelihood, are indicative of the adversarial effect of liquid nitrogen used during curettage.

The modes of reconstruction following wide resection of the DER can be broadly classified into either mobile joint (AP) or fixed joint (AD). Additionally, the choice of reconstruction material can also be highly variable. Because of near identical geometries and similarity of joint surface, the proximal fibula is often used as a substitute for distal radius reconstruction.^18,33 Both vascularised and non-vascularised fibula transfers can be used. However, according to Vander Griend et al.,⁷ the usefulness of vascularised graft may be limited due to (i) short length of the graft (less than 9 cm) (ii) ability to cover at least part of the graft with extensor pollicis/extensor carpi ulnaris muscles, and (iii) use of a stable fixation construct which would protect the graft from revascularization. Besides fibula, the distal ulna can also be used as an alternative. However, ulnar translocation can produce significant narrowing of the distal forearm, creating an hour-glass appearance, which might not be cosmetically appealing to many patients.⁷ Osteoarticular allografts are another option for reconstruction. Since a distal radius allograft is size matched with the carpus, AP is usually performed in such cases. However, allografts are plagued by high incidence of non-union and limited availability causing limitation on their usage.^34,35

There are very few studies that have directly compared the two techniques of wrist reconstruction.^7,9,18,19,36 All of the studies have demonstrated better preservation of wrist ROM with AP. However, it is associated with a high index of instability and has a tendency to cause carpal subluxation.^18,37 On the other hand, not only does AD stabilize the wrist joint and improve grip but patients undergoing fusion also have reported a more favourable functional result.^9,18,19 Our review has shown acceptable results with AD in terms of both satisfactory pronation/supination and significantly lower complication rates. This coupled with the preserved motion of the small joints of the hand, compensate for the loss in wrist mobility associated with the procedure.^9,18 We therefore, believe that wrist AD would be the ideal procedure in young, active adults who require a stable, powerful wrist. AP, conversely, can be reserved for patients who desire more flexibility at the wrist and are not engaged in heavy, manual work.^7,9,18,19

Our review had a few limitations. Out of thirteen studies, only two were found to be prospective in nature. A meta-analysis of retrospective studies is undesirable given the predisposition to confounding and selection bias. However, due to the dearth of prospective studies or randomized trials pertaining to this topic, it was the only feasible option. Secondly, data analysis on functional outcome measures was limited due to the inconsistent reporting in a number of studies. Finally, due to the paucity of available literature there were very few studies meeting our inclusion criteria for the second half of the review.

Despite these limitations, all steps were taken to ensure methodological quality of this review. The study criteria were laid down as per established PICO format. Two databases were searched. Bias was minimised by having two independent reviewers screen the search results, with additional inputs from a third expert in case of any inter-observer differences. Reviewers carrying out ROB assessment were blinded to each other. To the best of our knowledge, this study is the first of its kind to make an objective comparison between the two modalities of wrist reconstruction following wide resection of distal radius GCTs.

In conclusion, this review highlights that there is a definite advantage of wide resection over curettage in terms of local recurrence rates, especially in Campanacci grade III tumours of the distal radius. However, curettage being a less morbid procedure provides better or equivalent functional outcomes with reduced complication rates. In terms of choice of reconstruction after wide resection, wrist arthrodesis provides improved grip strength and better overall patient satisfaction as compared to arthroplasty. However, further research in the form of well-designed prospective randomised controlled trials are needed to validate the best method of reconstruction following wide resection.

Source of funding

This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

CRediT authorship contribution statement

Divesh Jalan: Writing – review & editing, Supervision, Validation. Akshat Gupta: Writing – original draft, Conceptualization. Raghav Nayar: Writing – review & editing, Data curation, Formal analysis. Nupur Aggarwal: Writing – review & editing, Data curation, Formal analysis. Kuldeep Singh: Data curation, Methodology. Princi Jain: Data curation, Methodology.