Palliative treatment of painful bone metastases: Does fractionation matter?
Metastatic bone pain is acommonly encountered clinical condition seen in oncology clinical practice. About 50% of all cancer patients develop metastases in their lifetime and half of them develop skeletal metastases. Despite its importance, no specific therapeutic strategy, to prevent or treat this complication of cancer has been demonstrated. This study comparing two radiotherapy regimens was conducted at a cancer center in New Delhi, India. Patients with confirmed bone metastases were randomized into two treatment arms, Group A (single fraction) and Group B (multiple fractions). Patients were followed up for 12 weeks. Pain relief was the primary endpoint of the treatment. Other parameters were improvement in analgesic score, performance status, and acute side effects like nausea, vomiting, tiredness, and lassitude. Fifty patients were evaluated at the end of the study. Overall response rate was seen in 86% of cases, whereas complete response was seen in 36% of cases. The two treatment regimens were found to be comparable with respect to other endpoints. Hence, a single fraction treatment, which is more convenient and cost effective, is a more logical approach in the Indian scenario in selected cases.
Keywords: Bone, fractionation, metastases, pain, radiotherapy
Bone metastasis is a frequent condition encountered in oncology practice. Bones are the third commonest metastatic site after lungs and liver. Tumors leading to bone metastases include breast (50.1%), prostate (16.6%), and lung (11%). Other primaries include bladder, kidney, uterus, melanoma, thyroid tumors, and unknown primaries.  Pain is the most common symptom, present in 75% of patients. 
The incidence of bone metastases is increasing because of longer survival rates amongst cancer patients.  Bone metastasis is usually detected after diagnosis of primary but in 10-15% of cases they are the first lesions to be detected.  In 50-70% of patients, complications like pain, pathological fracture, hypercalcemia, nerve root damage, and cord compression are seen. Bone metastases are usually multiple. Solitary metastasis is produced only in <10% of cases. Most commonly affected areas are spine, pelvis, ribs, proximal thigh, and skull. 
The pathology of bone pain is poorly understood. Mechanisms that cause pain include structural damage, periosteal irritation, nerve entrapment, muscle spasm, and secretion of mediators like prostaglandins and cytokines, which activate osteoclasts and nociceptors.  There is no correlation between metastatic pain and the type of cancer, site, size, or number of metastases.  The mechanism of pain relief following radiotherapy (RT) is not completely understood. It may not be dependent on tumor shrinkage and cell kill only. Other possible mechanisms include an effect upon the release of chemicals mediating pain response like prostaglandins and neurogenic peptides from host cells.
X-rays play an important role in screening for metastatic bone lesions. Computed tomography (CT) scan or magnetic resonance imaging (MRI) are useful for assessing disease extent and for RT planning. MRI is good for assessment of marrow replacement by tumor and soft tissue extension while CT gives superior details of cortical bone destruction. 
Management of bone metastases presents a challenge to oncologists. Its current management includes RT, chemotherapy, hormone therapy, surgery, and supportive therapy either alone or in combination.  In most cases the treatment intent is palliative, when treatment goals are pain relief, preservation of mobility and function, preservation of quality of life and if possible, prolongation of survival. Only in exceptional cases is the treatment goal curative.  RT is the most effective treatment of bone metastases.  At least 75% of patients achieve pain relief following RT and about half of them stay free from pain.  However, there is no consensus regardingdoses or fractionations required to achieve these results.  A number of trials have been conducted in the past to answer the controversies regarding the most appropriate RT schedule but, there are marked difficulties in the interpretation of their results due to methodological deficiencies.  For example, in 1982, a trial comparing various schedules concluded that "low dose short course schedules are as effective as more aggressive protracted schedules" in controlling metastatic bone pain.  However, in 1985 the same data was re-analyzed using different techniques and it was concluded that "protracted schedules produced more pain relief than short course schedules."  A large number of trials have have shown no difference between rates of onset, duration and degree of pain relief among different schedules involving single or multiple fractions.  Despite evidence from trials, some authors still believe that a higher total dose provides better pain relief of longer duration than a single fraction. 
The measurement of pain has always been a problem for researchers and clinicians working in palliative studies. Although several scales are currently available, it remains unclear which of these provides the most precise measurement because of the subjective nature of this experience. Despite these hurdles, new research studies continue undeterred to elucidatethe exact role of RT in the management of painful bone metastases.
The purpose of this study was to compare single fraction RT with multiple fraction RT in the palliative treatment of painful bone metastases in Indian patients where metastatic disease constitute a significant proportion of our total cancer workload in RT departments, as >50% of the patients present in advanced stage disease and ultimately develop metastases. This study addresses a therapeutic question of considerable clinical significance.
The study was an open randomized trial conducted in the Department of Radiotherapy, Maulana Azad Medical College and Lok Nayak Hospital, New Delhi, India from February 2004 to August 2005.
The study included patients with painful bone metastases localized to a single region that could be encompassed in a single radiation field. In all patients, the malignancy was histopathologically confirmed and the metastases were radiologically verified. Patients with any primary, melanoma and thyroid carcinoma excepted, were included. Other inclusion criteria were presence of pain that could be evaluated during and after treatment, and willingness to sign the informed consent. Patient could re-enter the study, if they developed a separate, previously untreated, painful lesion requiring RT.
The exclusion criteria were previous RT to the region concerned, presence of any co-morbid condition to which the patient's symptoms could be attributed and inability to understand Hindi or English.
The study protocol was presented tothe ethics committees of the respective hospital and medical college and cleared .Patients fulfilling the inclusion criteria were randomized into two equal treatment groups, Group A (Gp-A) and Group B (Gp-B). Randomization was done using a table of random numbers. Informed consent was obtained from all the patients before they underwent clinical evaluation involving a detailed history, physical examination, baseline laboratory investigations, and imaging studies. Evaluation of baseline pain, analgesic consumption, performance status, and any other significant symptom was done and recorded on the first day of treatment. At the same time, the scoring system was explained to the patient.
Gp-A patients received 8 Gray (Gy) in single fraction and Gp-B patients received 15 Gy in five fractions. Both the groups were started on RT within a week of randomization. A single direct field or two parallel opposed fields were used as dicatated by the needs of each case.In general, the policy was to use a field encompassing the painful areas with a generous margin. Supplementary treatment for relieving distressing symptoms was given on an as and when required basis. Treatment with chemotherapy and/or hormonal therapy was allowed, but all changes in the treatment were registered.
After treatment completion patients were followed up weekly for 4 weeks and then bi-weekly till the end of follow up at 12 weeks. Patients were followed up in person at the clinic on the first, fourth, and twelfth week. On other occasions, patients were reviewed via telephone. At each follow-up, assessment looked at the patient's general condition, performance status, pain intensity, analgesic consumption, and side effects. Day 1 was taken as the day on which the first fraction of radiation was delivered. Patients who developed new painful metastases remained on-study and these events were treated appropriately.
The pain score was evaluated on a 11-point scale. It consisted of numbers from 0 to 10 where 0 represented no pain at all and 10 represented the worst pain one could imagine. The score of 1-9 represented variable severity of pain between the two extremes. The patients continuously recorded their analgesic intake. Analgesic intake and acute side effects like nausea, vomiting, and lassitude were evaluated on 0 (none) to 4 (maximum) scale. Performance status was evaluated according to the Eastern Co-operative Oncology Groups (ECOG) criterion.
Nausea and vomiting scale
0: No nausea or vomiting
1: Nausea but no vomiting
2: Transient vomiting not requiring anti-emetics
3: Vomiting requiring anti-emetics
4: Intractable vomiting
Tiredness and lassitude scale
1: Slight tiredness
2: Tired and lethargic
3: Lethargy requiring bed rest
4: Extreme lassitude
Analgesic requirement scale
1: Simple analgesics (aspirin, ibuprofen)
2: Mild narcotics (codeine, tramadol)
3: Strong narcotics (morphine, fentanyl)
4: High-dose narcotics inadequate
The extent of pain relief was the main indicator for effective palliation .
Response was defined as improvement in pain score by at least two points with respect to the pretreatment value.
Complete response was defined as achieving a pain score of 0 at any point during follow-up.
Time for onset of response was the day when patient reported improvement in pain score by at least two points.
Duration of response was defined as time from initial response till return of pain to its baseline value.
Duration of complete response was calculated from date of complete pain relief to the date of increase in pain above 0.
Progression was defined as an increase in pain score above the baseline score without any response or after achieving initial response.
Descriptive statisticswere calculated for the parameters in the study, whenever applicable.
Continuous and ordinal variables were subjected to the Student t-test, and Wilcoxon's ransom test while categorical data were analysed using theChi-square test.Two-way ANOVA (Friedman nonparametric) was performed to observe the trend within the variable. P-value of <0.05 has been considered as statistically significant. SAS 8.0 was used to analyze data.
Observation and results
Overall patient profile
Out of 56 enrolled patients, 50 completed their follow-up and were evaluated at the time of data compilation [Table 1]. The main reasons for exclusion included having received RT earlier, multiple painful sites, patients having painless bone metastases, a pathological fracture, or a combination of these factors.
In Gp-A, patient's age ranged from 25 to 70 years with a mean of 58.48 years. In Gp-B, patient's age ranged from 28 to 70 years with mean of 53.28 years. Majority of patients, i.e., 9 (36%) in Gp-A and 10 (40%) in Gp-B were in range of 41--50 years. No significant difference in age distribution was observed between two groups (P = 0.69). Most of the male patients reported some form of substance abuse. Tobacco was the commonest substance of abuse in both groups [Table 1] and [Table 2].
Majority of the cases in both groups, i.e., 7 (28%) in Gp-A and 6 (24%) in Gp-B had lung cancer. All patients presented with pain. Other symptoms like restriction of movements, symptoms due to sensory motor deficit and autonomic disturbance, and loss of weight and appetite were also present in many patients. In majority of cases, the lumbo-sacral spine was involved. Beside this the pelvis was the commonest site. Other involved areas were ribs, shoulder girdle and skull. Overall, spine was involved in 27 (54%) of the total 50 patients.
In majority of the patients, multiple bony sites were involved. In 12 (48%) patients of Gp-A and 12 (48%) of Gp-B, primary was controlled while patients were having symptomatic distant disease. In others, primary was either uncontrolled or the status was unknown. Most of the patients were not receiving any systemic treatment during RT or follow-up while others were receiving it. However, the distribution was not statistically significant between two groups (P = 0.89) [Table 2].
Overall response to treatment
In both the groups, mean pain scores improved from week 0 to 12, which indicated adequatepalliation as defined for the pruposes of this study. Majority of the patients, i.e. 43 (86%) of the total 50 patients obtained overall pain relief following RT. Out of them, 18 (36%) obtained complete pain relief. Seven patients (14%) did not report any benefit. There was no statistical difference between the two schedules (P = 0.68) although the multiple fractionation group has shown a slight advantage in pain relief.
In 27 (54%) of the total 50 patients, there was a decrease in the analgesic consumption following RT. This improvement was maintained till the end of follow-up in both groups. In 17 (34%) of the total patients, there was no change in the analgesic requirement. Both the treatment groups were similar in this respect [Table 3].
In majority of the patients, i.e. 36 (72%) out of 50 patients, the performance status did not change from the baseline value. In only 6 (12%) patients, it improved during the follow-up. The trend was similar in both groups.
Overall, 27 (56%) of the total 50 patients did not report any adverse effect during follow-up. Although more adverse effects were reported in Gp-B, the difference was not statistically significant (P = 0.77). The only significant difference observed during the first week involved nausea and vomiting (P = 0.0486).
Duration of various responses to treatment
Onset of response
Majority of the responders, i.e. 26 (60%) out of a total 43, reported response within the first week. Others reported improvement in second or third week. There were only 4% patients reporting it after the third week. There was no significant difference between the two groups regarding onset of response (P = 0.73) [Table 4].
Duration of overall Response
Majority, i.e., 23 (53%) of total 43 responders maintained the response for more than 10 weeks or till the end of follow-up. Shorter duration of responses were recorded mainly because of intervention by death. The difference was not statistically significant between two groups (P = 0.80).
Duration of complete response
Complete response was maintained for a shorter duration because of limited follow-up. In majority of complete responders, i.e., 7 (38%) out of 18, it was <3 weeks. However, it was similar in both the groups.
The mean and median overall survival was comparable in both the groups (P = 0.85) [Table 1]. This can be seen in Kaplan-Mier curve [Figure 1]. When survival was analyzed according to primary site, it was seen that patients with carcinoma breast experienced longest survival while patients with lung cancer had shortest survival durations.
Ten patients from each group underwent radiologic evaluation with CT or MRI. Majority of the patients were found to have lytic lesions, followed by mixed lesions. Sclerotic lesions were very few in both the groups. In the majority of patients, no change was observed in the lesion after completion of follow-up. Only two patients showed evidence of regression in the post-treatment CT scan [Table 5].
Metastatic bone disease is the most common malignant bone lesion seen in adults. The management of these patients is nearly always palliative, but as such requires no less judgmental skill, inter-disciplinary co-operation, and effort than does curative treatment. The orthopedic surgeon and medical oncologist along with the radiation oncologist, must be aware of the most appropriate measures and their timing during management.
In the present study, 82% of patients obtained pain relief and more than three-fourth of them obtained a 50% pain reduction, whereas 14% did not obtain any pain relief. This response is entirely in accordance with the findings of published trials. ,, As in literature,  most of our responders reported onset of improvement within the first 3 weeks. Otherwise, they did not responded to treatment in majority of cases. There were few late responders.
Patients obtaining complete pain relief (36% only) were less than those reported in literature. , This can be explained by relatively shorter follow-up. Very few responders could be followed till the time required for complete response to occur. Similar factors may be responsible for the shorter duration for which the complete response was maintained as compared to overall response.
There are several difficulties involved in this kind of trial and include patient selection, choice of end points, definition of precise contribution from RT and other co-interventions like analgesics, systemic treatment, and psycho-social support. In the present study, systemic treatment was allowed for ethical reasons.
Investigations involving measurement of pain relief are difficult to analyze because some patients do not survive long enough to complete full assessment.  In our study, this was seen in 25% of cases who died before completion of follow-up. Other major problem is measurement of pain. Several scales like the visual analog scale (VAS), 101-point scale, 6-point behavioral scale, 4-point verbal rating scale, amongst others have been documented in current literatured.  In our study, the 11-point scale was found to be a convenient for patients as it is simple to understand and provides an adequate range to choose and report. Moreover, it is easy administered over thetelephone. The compliance of the patient has always been a problem in studies involving palliative treatment because patients usually have advanced disease and poor life expectancy. Moreover, the care givers often lacked the motivation to assist with follow-up. In countries like India, the noncompliance rate is very high because a large number of patients switch over to alternative therapies like ayurveda, yoga, and other local therapies due to the lack of expected relief. The compliance of the patients was better than expected in our study as the number of hospital visits were kept to a minimum and all the patients were followed up via telephone to update their score forms. Most patients completed their forms in time. There was no loss to follow-up. Only during the final evaluation, six patients were excluded because they did not fulfill the desired criteria. This method of follow-up seems logical and can be used in future studies for terminally ill cancer patients who have to travel long distances to hospital in developing countries where optimum cancer care facilities are available only in major hospitals in bigger cities.
Exsisting literature indicates no correlation between the type of primary and the likelihood of pain relief. , However,the present study indicates a poor response to pain relief in patients with renal cell carcinoma. Only one out of three patients with renal cell carcinoma showed some response to the intervention studied and this primary contributed half of the nonresponders. This observation needs further analysis in future as this study was under powered to draw any definitive conclusions.
Current available evidence onreported adverse effects is sparse, but there were no significant differences between the fraction schedules studied as far as the incidence of nausea, vomiting, diarrhea, and pathological fractures were concerned. 
In our study, more side effects were seen in the multiple fractionation group, but the difference was not statistically significant. It could be related to the longer stay (5 days) for these patients in hospital during RT and thus better reporting. A self-limiting increase in pain score was seen just after starting treatment in many patients, especially with vertebral metastases. This can be attributed to radiation-induced edema leading to increase in pressure over nerves. The incidence of this phenomenon was similar in both groups.
Published trials have reported the incidence of events like pathological fracture and spinal cord compression. None of our patient reported similar bone events during the 12 weeks of follow-up. However, one patient in Gp-A reported cord compression and two patients in Gp-B reported fracture during extended follow-up.
Another subject that generates interest is the issue regarding re-irradiation of the index site. Few trials , have reported higher rates of re-irradiation in the single fraction arms. An explanation for this observation may be the eagerness on the part of radiation oncologist to re-treat those patients who have received a lower dose in a single fraction arm earlier. In this study, one patient in Gp-A and two patients in Gp-B were re-treated, but both recieved these additional interventions after the cessation of follow-up 12 weeks later. We did not evaluate pain relief after re-irradiation, but others have shown that patients who have responded initially will have a similar probability of response after re-irradiation.  Further studies with larger sample sizes and a longer duration of follow-up are required to validate this issue.
In majority of patients who underwent CT or MRI, the lesions were purely lytic in nature. When the post-treatment (after 12 weeks) lesion was compared with pretreatment lesion, no change was reported in majority (80%) of them. Only two cases have shown evidence of healing by development of sclerotic rim around the lesion. Although, there are studies which document radiological responses within 12 weeks of treatment,  this duration of follow-up appears short for evaluating response. Our study is limited by small sample size, but confirms that it is feasible to evaluate re-mineralization of osteolytic lesions with palliative RT. In all cases, clinical response was seen much earlier than radiological response. Similar observations have been recorded in various trials.
According to the present study, both the treatment schedules are equally effective with respect to the evaluated variables and no regimen is superior to other. When pain relief is the primary goal, as it is in most cases of uncomplicated painful bone metastases, treatment with single fraction regimen may be more appropriate. When treatment objectives other than pain relief are equally important, the choice of schedule require further consideration. There may be patients with a solitary metastatic lesion in the skeleton or patients with a lesion in weight bearing bones who may benefit from a higher total dose with altered fractionation.
There is a need for further studies involving larger numbers of patients with longer durations of follow-up. Experiences with this study should inform the design of future studies to overcome currently identified deficits with this study.
We are of the opinioin that single fraction schedules may be preferred for patient convenience without compromising the palliative effect. The use of a single fraction could be of benefit to hospital staff and treating institutions especially in the Indian scenario where the radiation therapy departments are already overburdened. It also allows for more chances of re-treatment as the maximum tolerated dose of the surrounding structures is not crossed
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]