TABLE 2.

Cox proportional hazards regression analysis to determine factors associated with post-recurrence survival in 18 patients with histologically proven local relapse of extremity osteosarcoma after limb-sparing surgery

Factor	N	Post-recurrence survival (years)
		Median (range)	P-value
Demographics
Gender
Female	8	1.03 (0.31–8.02)	0.7060
Male	10	0.99 (0.40–12.11)
Race
White	13	0.85 (0.31–12.11)	0.7495
Other	5	1.25 (0.72–3.63)
Primary tumor
Site of primary tumor
Upper extremity	4	0.68 (0.40–1.78)	0.3924
Lower extremity	14	1.17 (0.31–12.11)
Necrosis of primary tumor
Rosen grade 1 and 2	11	1.25 (0.31–12.11)	0.1447
Rosen grade 3	7	0.81 (0.40–3.63)
Distant metastases
Metastases at initial diagnosis
Present	10	0.99 (0.31–8.02)	0.7747
Absent	8	1.05 (0.34–12.11)
Metastases at time of local relapse
None or completely resected	16	1.20 (0.30–12.10)	0.0362a
Not resected	2	0.55 (0.30–0.80)
Metastases after local relapse
None or completely resected	9	1.80 (0.30–12.10)	0.3234
Not resected	9	0.90 (0.40–5.40)
Metastases at any time in treatment course
None or completely resected with no further distant recurrence	8	2.70 (0.30–12.10)	0.1289
Not resected	10	0.85 (0.30–5.40)
Local recurrence
Type of resection performed for local recurrence
Amputation	10	2.44 (0.30–12.10)	0.2551
Limb-sparing wide excision	6	0.86 (0.34–1.78)
No surgeryb	2	0.45 (0.42–0.48)
Surgical margins of resection of local recurrence
<1 cm or positive margins	6	0.65 (0.30–1.30)	0.0277
≥1 cm	10	2.70 (0.70–12.10)
No surgeryb	2	0.45 (0.40–0.50)
Time from diagnosis to local recurrence		1.40 (0.62–10.37)	0.1565

^aP-value may be exaggerated due to small numbers in a category.

^bPatients who did not undergo surgery were excluded from analysis because of small numbers in this category.