¹⁸⁶Re-HEDP in the Treatment of Patients with Inoperable Osteosarcoma

Patients

Between March 1999 and December 2004, 13 patients (9 male, 4 female) with histologically proven osteosarcoma and a positive ^99mTc-HEDP bone scan result (i.e., showing higher tracer uptake in tumor than in normal bone) were included in the study. The median age of the subjects was 16 y (range, 12–42 y), and their average weight was 52.5 kg (Table 1). All patients consented to receive the therapy. The intent in all patients was palliative.

Of these 13 patients (Table 1), at presentation 5 had primary axial tumors without evidence of metastatic disease. These tumors were inoperable, and ¹⁸⁶Re-HEDP therapy was part of the planned initial management strategy to supplement radical external-beam radiation to the primary tumor. All patients had been treated with a standard neoadjuvant chemotherapy protocol, had a clear response after 2 cycles, and had a further response or stable disease after a total of 4 cycles.

One patient who had had an isolated pelvic bone metastasis was treated in the same manner after second-line chemotherapy, receiving a single administration of ¹⁸⁶Re-HEDP immediately before 55-Gy EBRT. The remaining patients had established metastatic disease and received ¹⁸⁶Re-HEDP after second-line or further chemotherapy with the aim of palliation or maintaining stable disease (Table 2). Eight of these patients had originally been treated with combination chemotherapy followed by surgery of the primary tumor (amputation in 4, limb salvage surgery in 3, or radical radiotherapy in 1). Relapse had occurred at a bony site in only 2 patients and in soft tissue in 6; of the latter, 2 had had previous pulmonary metastasectomies.

All patients underwent a full disease review, a general physical examination, and a routine blood count, liver function tests, glomerular filtration rate test, and measurement of left ventricular ejection fraction before EBRT and radionuclide therapy.

Therapy Protocol

The ¹⁸⁶Re-HEDP was purchased from a commercial source (Mallinckrodt Medical Ltd.). Patients were treated as inpatients with an ¹⁸⁶Re-HEDP activity of 18.5 MBq/kg. ¹⁸⁶Re-HEDP was administered by slow-bolus (approximately 15 s) intravenous injection, with close monitoring of vital signs. Posttherapy scans were obtained at 3–4, 24, 48, and 96 h. Posttherapy response was assessed using the criteria of the World Health Organization (11).

Criteria for Rhenium Administration

Dosimetry

Tumor dosimetry was performed using the formalism developed by the MIRD Committee (12–15). Adopting the notation of the MIRD schema (16,17), the absorbed dose D(t←s) to target region t from source region s is:Eq. 1where Ã_s is the cumulated activity (i.e., the total number of nuclear disintegrations) in source region s and S(t←s) is the source region s–to–target region t S factor (i.e., the mean absorbed dose to target region t per nuclear disintegration in source region s). This formalism permits calculation of the mean absorbed radiation dose delivered to a target region from radioactivity distributed within one or more source regions within the body, and where the source and target may be the same region. In this latter case, it is the “self-irradiation” contribution of the total dose delivered to the target region that is determined, where this is usually the predominant component. In its usual application to diagnostic radiopharmaceuticals, the MIRD schema assumes that the activities and cumulated activities are uniformly distributed within source regions and the radiation energy is uniformly deposited within target regions defined. Thus, if the MIRD formalism is used at the organ level, it is the mean radiation dose within the target organ that is calculated. Use of the MIRD schema has been adapted here to determine the mean absorbed dose delivered to a tumor assumed to be homogeneous in composition, and confined to the component of the total dose that is due to “self-irradiation” by activity distribution within the tumor region itself.

Determination of Tumor Activity from Image Data

Conjugate anterior and posterior whole-body imaging was performed at 0, 4, 24, and typically 96 h after injection using a dual-detector (Maxxus; GE Healthcare) large-field-of-view γ-camera with low-energy high-resolution collimators. The 0-h image dataset was acquired before the patient was allowed to micturate and thus reflected the in vivo retention of all administered activity. Imaging was initially performed at a 15 min/m scan speed, which was modified at later time points to compensate for physical decay of the tracer, and all image counts were adjusted for differences in scan speed.

Irregular regions of interest were defined to encompass the tumor boundaries, generated from the initial image dataset, and overlaid on all subsequent images, correcting for patient placement at each time point. These regions were drawn in direct reference to those placed on the contiguous (transaxial)-slice CT or MRI data, and all possible attempts were made to ensure that these matched as closely as possible. Partial-thickness background correction was also applied using a technique appropriate for a single, well-defined source region surrounded by background activity (18) by using representative peripheral irregular background regions of interest and an estimate of fractional tumor thickness if the tumor did not involve the full anteroposterior body thickness at that site. Geometric mean tumor region-of-interest count data were then expressed as the fractional uptake of the geometric mean whole-body counts for the 0-h image dataset—where this is taken to represent 100% of the administered activity in vivo.

Tumor time–activity curves were then plotted from corrected fractional tumor region-of-interest count data, and a single exponential curve was fitted and, if necessary, weighted to the later image time points to better reflect retention of the tracer after the initial phase and physical decay of the tracer. Although this strategy introduces some inaccuracy in the estimation of cumulated tumor activity (represented by the integrated area under the curve), this is a pragmatic strategy in response to the necessity of working with the relatively few image data points gathered for these very sick patients, returning as outpatients after the first 24 h after injection. It is probably reasonable to assume an effectively near-instantaneous uptake of tracer into the osteosarcomatous lesion.

The tumor time–activity data, that is, the non–decay-corrected tumor activity (in MBq), were fit to a monoexponential function:Eq. 2where a is the zero-time intercept (in MBq), b the effective clearance constant (in h⁻¹) of the exponential function, and t the time after administration (in h). Integration of this function and normalization to the administered activity yields the tumor cumulated activity per unit administered activity, Ã (in MBq-h/MBq):Eq. 3where A_o is the administered activity (in MBq).

Estimation of Mean Tumor Dose

The cumulated activity determined for the tumor above was normalized to the net administered activity (corrected for residual activity within the injection syringe) to obtain the actual cumulated activity for this administration. This figure was then multiplied by the S factor determined below to yield the estimated absorbed dose to the total tumor.

Determination of Tumor Volume, Mass, and S Factor

Contemporaneous or near-contemporaneous CT data were obtained for the tumor site for all but 1 of the patients studied, and with the exception of 2 further patients for whom contemporaneous MRI data of the tumor site were available. Image data were obtained in film format only, necessitating a manual approach to the accurate determination of tumor volume and composition. All image data were reviewed critically, and the boundaries of the tumor were transcribed directly to tracing paper for each contiguous slice. At the same time, this slice was manually segmented into regions determined visually from the underlying CT data as comprising either bone, calcified soft-tissue involvement, “sparsely” calcified soft tissue, or soft tissue, as appropriate; some tumors were deemed to contain only 1 of these components, some 2, and a few all 4. External tumor boundaries were delineated in relation to the limiting functional tumor volume defined by tracer uptake observed from the scintigraphic data.

The total volume represented by each of the 4 tissue components was determined by summation of the traced regions for all slices, using grid-square graph paper as a scaling device. The displayed CT or MR image and knowledge of CT or MRI slice thickness (and pitch data for spiral CT protocols) were then used to normalize this summed area in arbitrary units of “graphical squares” to the true physical size of each tumor region (in cm³).

Each region was then allocated an appropriate mean tissue density for estimation of its total mass. This was determined as ranging from a tissue density of 1.99 g cm⁻³ for cortical bone to 1.04 g cm⁻³ for soft tissue (taken as skeletal muscle). Tissue regions defined as calcified and sparsely calcified were allocated an intermediate mean density, determined by an assumption of their fractional components by volume. For example, 1 tumor was deemed to have a component of calcified soft tissue best represented, assuming it consisted of 40% cortical bone and 60% soft tissue. Thus, the density in this case was determined to be 1.42 g cm⁻³ (0.4 × 1.99 g cm⁻³ + 0.6 × 1.04 g cm⁻³). A further component of sparsely calcified soft tissue was identified to be present within some tumors, where this was likewise assessed by inspection of the CT Hounsfield unit values within the tumor to comprise relative portions assignable to bone and soft tissue reflecting 10% islands of cortical bone within an environment of 90% soft tissue, thus yielding a mean density of 1.14 g cm⁻³ (0.1 × 1.99 g cm⁻³ + 0.9 × 1.04 g cm⁻³).

The total mass of each tissue component was then determined from the product of its assigned mean tissue density and the volume estimated from the traced limits defined from the CT data. The total mass of the whole tumor was then calculated.

The appropriate S factor for ¹⁸⁶Re for this tumor mass was then determined from the interpolation of a set of S factors for “idealized spheric tumors” accessible through the MIRDOSE3.1 dosimetry software package (19). A log–log plot was generated for the log₁₀ S factor for ¹⁸⁶Re (in mGy/MBq·s) (y-axis) versus the log₁₀ mass of an idealized tumor sphere (in g) (x-axis), for which a relationship of y = −0.9760x – 1.3218 with a regression of R² = 0.9997 was obtained. This relationship was then used to determine the S factor for the estimated total tumor mass. The assumption of a spheric geometry does not in itself introduce a significant error, given the relatively short range of the β-particles emitted by ¹⁸⁶Re in tissue and their consequent insensitivity to tumor morphology (20).

Treatment-Associated Toxicity

Two patients experienced white blood cell and platelet toxicity of more than grade 3 during or after ¹⁸⁶Re-HEDP therapy (Table 3). Patient 6 experienced grade 4 platelet toxicity and required a transfusion. For personal reasons, this patient received EBRT to a large pelvic field within 2 mo of ¹⁸⁶Re-HEDP therapy, possibly contributing to the marrow toxicity. There were no episodes of neutropenic sepsis. There was, however, evidence of an effect on bone marrow function. In patient 3, administration of palliative oral etoposide, 2 mo after a fourth administration of ¹⁸⁶Re-HEDP, produced profound marrow depression, a complication that would not have been expected as a consequence of previous chemotherapy or in the absence of evidence of marrow infiltration by tumor. Poor hematologic tolerance and gastrointestinal bleeding contributed to the death of this patient (Tables 2 and 3).

No significant acute or late bowel toxicity was observed in the 4 patients who underwent EBRT to the pelvis. Although only 1 of these patients has sufficient follow-up for accurate recording of neurologic toxicity, no cauda equina or sacral nerve root toxicity was observed, despite the fact that these structures were within the target volumes of EBRT. There was no deterioration in glomerular filtration rate or left ventricular ejection fraction after EBRT and ¹⁸⁶Re-HEDP therapy (P > 0.05). In patient 1, who was in chronic renal failure, the glomerular filtration rate was low (46 mL/min) before ¹⁸⁶Re-HEDP treatment because of ifosfamide therapy and showed no significant change after treatment.

¹⁸⁶Re-HEDP Therapy

The purpose of ¹⁸⁶Re-HEDP therapy was to supplement EBRT, to provide palliation to multiple sites, to consolidate the chemotherapy response, and to prolong the duration of stable disease. The single therapeutic activity used for ¹⁸⁶Re-HEDP was 18.5 MBq/kg (range, 1,182–5,310 MBq [72−143 mCi]). Of the 13 patients, 10 received a single administration of ¹⁸⁶Re-HEDP (1,182–1,500 MBq [32–40.5 mCi]). Two patients received 2 administrations (cumulative activities of 2,660 and 2,820 MBq) of ¹⁸⁶Re-HEDP, with a 6-wk interval between them (Table 1). One patient with bone metastases received 4 administrations of ¹⁸⁶Re-HEDP (cumulative activity, 5,310 MBq) (Table 1). Mean uptake of ¹⁸⁶Re-HEDP by the tumor was 5.8% of the administered activity. No acute side effects were observed after ¹⁸⁶Re-HEDP therapy.

Follow-up

Of the 3 patients in whom ¹⁸⁶Re-HEDP therapy was used to boost EBRT, 1 is alive with local control 8 y from therapy. The remaining 2 had local progression (Table 2). Of the 11 patients who died, 7 did so because of disease progression, 1 because of intestinal obstruction, and 1 because of intestinal perforation related to intraabdominal disease. Two patients had hemoptysis secondary to lung metastases, and their symptoms abated after EBRT. Seven patients experienced moderate to severe pain, either because of local progression of disease with tumor compression of nerve roots or local expansion of tumor. Overall, partial symptomatic response was observed in 50% and partial tumor response in 30% of patients. One patient received 4 administrations of ¹⁸⁶Re-HEDP, and stability was achieved for 12 mo before progression. The remaining patients progressed rapidly after the first ¹⁸⁶Re-HEDP therapy and were not considered suitable for further therapy. Of the 13 patients, 1 is still under follow-up and 1 has been lost to follow-up. No patient had complete remission of disease. The median survival time from the diagnosis of metastatic disease was 36 mo (range, 12–216 mo) (Table 2). The median survival time after the last therapeutic dose of ¹⁸⁶Re HEDP was 5 mo (range, 1–60 mo) (Table 2). The overall survival rate since the last therapy was 15% at 3 y.