Growth in the Lower Limb Following Chemotherapy for a Malignant Primary Bone Tumour: A Straight-Line Graph

Purpose. The aim of this paper was to assess the growth in the unaffected lower limb of children who had received chemotherapy for a malignant primary bone tumour around the knee. Subjects/methods. Following diagnosis, all children (45, of which 32 were boys and 13 were girls) were staged. If limb-salvage surgery was thought appropriate, measured radiographs of both legs was performed, the bone age was estimated and the expected growth in the femur and tibia was calculated according to Tupman. These procedures were repeated at follow-up and the data plotted. Regression and correlation coefficients were also calculated. Results. The observed regression line in boys was almost identical to Tupman's curve. However, the observed growth in girls was larger than the expected growth. Discussion. It is recommended that the regression lines presented here are used in the calculation of the expected growth in the lower limb of children who have received chemotherapy for a malignant primary bone tumour, especially in girls.


Introduction
N eo-adjuvant chemotherapy has greatly im proved the survival in children w ith m alignant prim ary bone tum ours. T he overall survival in patients with osteosarcom a is now adays 64%. 1 Im provem ents in diagnostic im aging, surgical technique, biom edical engineering and survival have m ade limb-salvag e surgery in children possible. One lim b-salva ge option for children w ith a m alignant prim ary bone tum our around the knee is replacem ent with a custom -m ade extensible endoprosthetic replacem ent. 2± 6 Our centre has perform ed extensible endoprosthetic replacements for prim ary bone tum ours around the knee since 1976.
In order to m anufacture the extensible endoprosthetic replacement, it is necessary to have an estim ate of the expected growth in the low er lim b. Our centre has used data provided by Tupman 7 to calculate this growth. However, these data are based on a norm al population of children.
Fraum eni 8 reported that children with osteosarcom a are taller at the time of diagnosis than controls. However, others 9,10 did not ® nd such a relation.
Glasser et al. 9 showed that children with m alignant prim ary bone tumours had a m arked retar-dation in growth during the year of cytotoxic chem otherapy. T hey concluded that ® nal height m ight be affe cted, but to a small degree. W e could ® nd no reports in the literature providing inform ation about the growth in the lower lim b of children w ho have received chem otherapy. T he aim of this study was to assess the growth in the lower lim b of children w ho had chem otherapy for a m alignant prim ary bone tum our.

Subjects and m ethods
Following the diagnosis of a m alignant prim ary bone tum our of the distal fem ur or proxim al tibia, all children were fully staged. T his included a radiograph of the lesion, m agnetic resonance im aging (M RI) or com puted tomography (C T) of the lesion, a bone scan and a CT scan of the chest.
If the patient was thought to be suitable for lim b-salva ge surgery, measured radiographs of both legs w ere also perform ed. The bone age w as also estim ated according to the Greulich and Pyle m ethod. 11 W ith the bone age, the expected growth in the fem ur and tibia was calculated according to T upm an. 7 All patients received chem otherapy in a neoadjuvant setting according to the then current  protocol. N o patients received radiotherapy to the extremities. Lim b-salva ge surgery w as perform ed and the patients w ere followed up in the outpatient clinic.
At the time of review , all patients had m easured radiographs of both legs and their bone age estim ated. The growth in the unaffected (norm al) leg w as calculated. As in Tupm an' s 7 review, the fem ur w as m easured from the top of the fem oral head to the m edial condyle. T he tibia w as m easured from the top of the tibial spine to the tip of the m edial m alleolus.
Patients who w ere skeletally im m ature at review had the expected growth at the time of review added to the m easured growth.
The data were plotted on a graph and superim posed on T upm an' s 7 curve. T he regression coef® cients were calculated by m eans of the least square m ethod and the regression line w as plotted in the ® gure. C orrelation coef® cients were also calculated.
In order to plan lim b-salva ge surgery, it is extremely helpful to have an estim ate of the expected growth in the proxim al and distal physis of the affected bone. T he growth in the proximal fem oral physis has been reported 7 as 30% and the growth in the distal fem oral physis as 70% of the total fem oral growth. Sim ilarly, the proxim al tibial growth is 55% and the distal tibial growth 45% of the total growth in that bone. 7 W ith the aid of these ® gures, regression lines for the proximal and distal physeal growth in the fem ur and tibia w ere also calculated and added to the ® gure.

Results
Between 1976 and 1992, our centre performed 106 extensible endoprosthetic replacements in children w ith a m alignant prim ary bone tum our around the knee. Of these patients, 39 died.
Patients were excluded if the follow-up was less than 1.5 years or if insuf® cient radiographs were availab le to assess the grow th. T hese criteria left 45 children in the study. There were 32 boys and 13 girls. The chronological age at diagnosis was on average 11.2 years (range 6.5± 15.3 years) and the m ean bone age was 10.6 years (range 6± 14 years). All patients were fully staged as described earlier and received neo-adjuvant chemotherapy according to the relevant treatment protocol. In 37 cases the diagnosis was osteosarcom a, in seven cases, Ewing' s sarcom a and m alignant ® brous histiocytoma in the rem aining patient. There were 29 patients who had a distal fem oral tum our. The rem aining 16 patients had a proxim al tibial tum our. The left side was affe cted in 23 cases and the right side in 22 cases. T he average follow-up was 4.5 years in both boys (range 2.1± 7.9 years) and girls (range 1.7± 11.1 years). O n average, the chronological age at followup was 16.0 years in boys (range 12.4± 20.8 years) and 14.7 years in girls (range 11.1± 22.8 years). Figure 1 show s the femoral and tibial growth that had occurred in boys and girls at the latest follow-up plotted against the bone age at diagnosis. T here w ere two boys who had identical m easurem ents, w hich explains why only 31 points are visible in the graphs for boys.
There were 17 boys (53%) and seven girls (54% ) w ho were skeletally m ature at latest time of followup (indicated by 3 1 ); the rem aining 21 children were skeletally im mature (indicated by s ).
The regression lines and correlation coef® cients are show n in the ® gure. F urtherm ore, regression lines for the proxim al and distal growth in the fem ur and tibia w ere calculated as described earlier and added to the ® gure. T upm an' s 7 original curve is also shown.

D iscussion
T he observed regression line in boys is alm ost identical to T upm an' s 7 curve. How ever, in girls the observed growth is larger than the expected growth. Although there were considerably less girls than boys in this study, virtually all girls had a grow th that was equal to or larger than expected. Furtherm ore, the correlation coef® cients are high. It seems therefore likely that the regression line in girls is a realistic indicator of growth. T he reason for the sex difference rem ains unclear.
Tupman 7 perform ed his study in 1962. T herefore, it is possible that his data are out of date. U nfortunately, m ore recent data are not availab le. H owever, com pariso n w ith our regression lines indicates that growth following chem otherapy is independent of skeletal m aturity prior to chem otherapy treatment. Furtherm ore, there does not seem to be a longlasting effect of chem otherapy on longitudinal growth in the low er lim b.
A longer follow-up time is unlikely to in¯uence the observed regression lines signi® cantly, since 53% of the children w ere skeletally m ature at the time of review and the correlation coef® cients are high.
In conclusion, we recom m end that these new regression lines are used in the calculation of the expected growth in the lower lim b in children w ho received cytotoxic chemotherapy for a m alignant prim ary bone tum our. This seem s especially im portant in girls.