The first generation of extendible endoprostheses appeared in the late 1970s. They were invasive, and open lengthening procedures were required. Several subsequent modifications have yielded the endoprostheses capable of lengthening through less invasive methods [14]. The initial part of the first generation was lengthened by means of a ball-bearing mechanism. It was replaced by a system which required insertion of a “C” collar in 1988 [15]. The Lewis Expandable Adjustable Prosthesis (Dow Corning Wright Corporation, Arlington, Tennessee) was also introduced in 1983. It used a fixed stem with a screw extension mechanism that expanded the endoprosthesis [16]. These endoprostheses had a high failure rate over the time required for multiple lengthening procedures [6].

The second generation of extendible endoprostheses was minimally invasive. Lengthening was achieved with an elongating screw mechanism that was rotated via a small incision. The need for soft tissue dissection was dramatically reduced, and joint motion was better, even though they also required an open procedure [15]. In 1987, the minimally invasive Growing Kotz prosthesis was introduced [4] (Fig. 10.2).

Third-generation extendible endoprostheses were noninvasive; lengthening was achieved without open surgery, and the risk of complications due to repeated surgeries was minimized [4, 17, 18]. The noninvasive extendible endoprosthesis (Wright Medical Technology, Inc, Arlington, Tennessee), originally known as the Repiphysis or Phenix prosthesis (Phenix Medical, Paris, France) [17, 19], and the Stanmore noninvasive extendible prosthesis (Stanmore, Middlesex, UK) were also introduced by the use of an external electromagnetic or rotating magnetic field for lengthening. Once implanted, these noninvasive extendible endoprostheses require no further surgical procedure during lengthening, thus allowing maximum extension without additional surgery.

Osteosarcomas predominantly occur in the metaphyseal region close to the growth plate. The distal end of the femur is the most common site for an osteosarcoma in children [20, 21]. When major growth plates are completely removed during limb-salvage surgery, the growth of the affected limb is seized before skeletal maturity.

Skeletal maturity is expected by the age of 16 years in boys and 14 years in girls [22, 23]. The amount of bone growth that can be expected from each growth plate has been described [24]. Up to 67 % of the lower limb growth occurs around the knee (distal femur and proximal tibia) [25]. The distal femoral physis presumably contributes to approximately 10 mm of growth per year until skeletal maturity [22, 23]. Significant leg length discrepancy results in gait abnormalities, leg pain, and back pain [26]. Kenan et al. reported that leg lengthening was necessary in patients younger than 13 years of age at the time of limb-salvage surgery [27].

As for the humerus, 80 % of total growth occurs in the proximal growth plate [25]. Ayoub et al. [28] mentioned that the main problem of arm length discrepancy is cosmetic. Functionally, lengthening of the humerus gains less functional advantage in comparison to the femur or tibia.

The necessity for an extendible endoprosthesis is based on the estimated growth remaining in the resected growth plate, calculated using the growth charts [22] after an estimation of bone age based on interpretation of a radiograph of the hand [29]. However, accurate prediction of future growth often becomes problematic. Dominkus et al. reported that the patients in their series required a mean expansion that was 24.3 % larger than the expansion anticipated after careful preoperative assessment of growth potential [30]. Implementation of a permanent solution at the time of the initial surgery may therefore lead to a discrepancy of unforeseen proportions. For this reason, an extendible endoprosthesis should be applied for the pediatric patient when the estimated leg length discrepancy at skeletal maturity is expected to be more than 3 cm or when the arm length discrepancy will be more than 5 cm [14, 28].

Several investigators have reported on patient functional scores after extendible endoprostheses. It has been widely accepted that extendible endoprostheses result in superior satisfaction and functional scores compared with those of amputation. Prosthetic reconstruction is well received emotionally and cosmetically by patients and their parents. Furthermore, the clinical and functional outcome correlates to that reported for adult patients [27, 31–33].

This author reported that the final functional score for endoprosthetic reconstruction was 74 %, in spite of the frequent necessity for additional operations, such as limb-lengthening procedures and revisions [6]. Function in the endoprosthetic group was comparable to the study of Schindler, in which the Musculoskeletal Tumor Society (MSTS) functional score was 77 % in 12 children with Stanmore custom-made extendible distal femoral endoprostheses [34]. Ruggieri et al. [35] reported that the mean MSTS score was 79 % (range, 30–100 %) by the use of the second and third generation of extendible endoprostheses; the mean MSTS score was 81 % (range, 63–93 %) for the Kotz Growing prostheses, 78 % (range, 33–100 %) for the Repiphysis, and 79 % (range, 30–100 %) for the Stanmore prostheses, without a significant difference between the type of endoprostheses.

Emotional satisfaction is an important factor for improved quality of life for young patients with extendible endoprostheses [36]. The Pediatric Outcomes Data Collection Instrument (PODCI) was developed to measure body image, social acceptance, as well as physical function and satisfaction after treatment for musculoskeletal disorders specifically in the pediatric population [36]. Henderson et al. [37] showed high emotional satisfaction for pediatric patients with extendible endoprostheses according to PODCI. Overall, patients reported excellent perceptions of body image and physical attractiveness. In addition, frequent social interactions and no difficulty in making new friends were also found.

Complications related to surgery, such as infection, aseptic loosening, implant breakage, dislocation, and skin necrosis, are frequent.

This author reported that 45 % of the skeletally immature patients with endoprosthetic reconstruction had experienced complications at 5 years, and the complication rate increased to 69 % at 10 years [6].

Deep infection was the most common complication. In the previous literature, the rate of deep infection in the femur varies from 4 % (1 of 24 patients) in the series by Cool [24] to 40 % (2 of 5 patients) in Schiller’s series [4]. Comparable results have been reported by others [34, 38–40].

Aseptic loosening has been reported as a major problem [4, 6, 35, 41], accounting for 10 % (4 of 39 patients) of complications in the series by Ruggieri P [36] up to 50 % (6 of 12 patients) reported by Schindler [34]. Picardo et al. [40] reported only 1 of 44 patients (2 %) with aseptic loosening by the use of the Stanmore noninvasive extendible endoprosthesis. Cannon et al. [42] reported that in distal femoral and proximal tibial resections, the rate of failure was higher when >60 % of bone was removed.

Previous studies have reported revision rates from a minimum of 20 % (5 of 24 patients) in the series by Cool [24] up to 100 % (12 of 12 patients) by Schindler [34]. Comparable results have been reported by others [4, 15, 24, 35, 39]. Unwin reported that 105 patients with distal femoral extendible endoprostheses had a probability of survival of 80 % at 4 years 29. This author reported on the long-term results of endoprosthetic survival: 77 % at 5 years and 51 % at 10 years, respectively [6]. Recently, Ruggieri et al. showed comparable results of 32 primary endoprostheses: 78 % survival at 5 years and 66 % at 10 years, respectively [35].