Estimating the optimal utilization rates of radiotherapy for hematologic malignancies from a review of the evidence

Part II—Leukemia and myeloma

Authors

  • Carolyn Featherstone M.B.Ch.B.,

    Corresponding author
    1. Collaboration for Cancer Outcomes Research and Evaluation, Liverpool Hospital, Sydney, New South Wales, Australia
    Current affiliation:
    1. Beatson Oncology Centre, Glasgow, Scotland, United Kingdom
    • Beatson Oncology Centre, Western Infirmary, Dumbarton Road, Glasgow, G11 6NT
    Search for more papers by this author
    • Fax (011) 44 1412116356

  • Geoff Delaney M.B.B.S., M.D.,

    1. Collaboration for Cancer Outcomes Research and Evaluation, Liverpool Hospital, Sydney, New South Wales, Australia
    Search for more papers by this author
  • Susannah Jacob M.B.B.S., M.D.,

    1. Collaboration for Cancer Outcomes Research and Evaluation, Liverpool Hospital, Sydney, New South Wales, Australia
    Search for more papers by this author
  • Michael Barton M.B.B.S.

    1. Collaboration for Cancer Outcomes Research and Evaluation, Liverpool Hospital, Sydney, New South Wales, Australia
    Search for more papers by this author

  • See also companion article on pages 383–92, this issue.

Abstract

BACKGROUND

The objective of this study was to estimate the ideal proportion of new patients with leukemia and myeloma who should receive radiotherapy at some time during the course of their illness based on the best evidence.

METHODS

Available evidence of the efficacy of radiotherapy in most clinical situations for leukemia and myeloma was identified through extensive literature reviews and treatment guideline searches. Epidemiologic data concerning the distribution of types, disease stages, and other factors that influence the use of radiotherapy were identified. Decision trees were constructed to merge the evidence-based recommendations with the epidemiological data to calculate the optimal proportion of patients who should receive radiotherapy according to the best available evidence. Actual radiotherapy utilization rates also were identified.

RESULTS

The proportion of patients diagnosed with myeloma in Australia who should receive radiotherapy based on the evidence was 38%. There was wide variation in the proportion of patients who actually received radiotherapy for myeloma from 24% up to 55%. The recommended proportion of patients diagnosed with myeloma in Australia who, according to the best available evidence, should receive at least a single course of radiotherapy was 38%. The proportion of patients diagnosed in Australia with leukemia who should receive radiotherapy at some point in their management, according to the best available evidence, was calculated at 4%, which corresponded with actual practice.

CONCLUSIONS

Further research will be required to determine why more patients who are diagnosed with myeloma are not treated with radiotherapy. Cancer 2005. © 2004 American Cancer Society.

Radiotherapy is an essential cancer treatment that contributes to the cure and palliation of many patients who are diagnosed with leukemia and myeloma. The planning of efficient and equitable radiotherapy services for a population requires an estimate of the optimal utilization of radiotherapy. A benchmark of the adequacy of radiotherapy services provision is the proportion of patients with new diagnoses of cancer who receive at least one course of radiotherapy, which we define here as the radiotherapy utilization rate.

Radiotherapy utilization rates for all cancers vary substantially by geographic region,1–6 with utilization rates ranging from 20% to 55% of all new cancer diagnoses. We have developed an evidence-based model of the use of radiotherapy in patients diagnosed with all cancers reported to statutory cancer registries. This report outlines the calculation of the estimate of the optimal utilization rate of radiotherapy in the treatment of lymphoma from the best available clinical and epidemiological evidence and compares the evidence-based optimal utilization with the actual utilization rates. The objectives of this study were 1) to estimate the optimal proportion of all patients with myeloma and leukemia who, according to best available evidence, should receive at least one course of radiotherapy at some time during the course of their illness; and 2) to compare the optimal radiotherapy utilization rate with actual radiotherapy utilization from current practice.

MATERIALS AND METHODS

Indications for Radiotherapy

An indication for radiotherapy was defined as a clinical situation in which radiotherapy is recommended as the treatment of choice based on evidence showing that radiotherapy produces a superior clinical outcome compared with alternative treatment modalities (including no treatment) and in which the patient is suitable to undergo radiotherapy. The superiority of radiotherapy may be based on survival, local control, or toxicity profile. An indication for radiotherapy may occur either in the initial stages of treatment or on treatment of recurrence. Patients who required radiotherapy were counted only once, even if they had multiple indications for radiotherapy at different stages during their illness.

National level guidelines and guidelines issued by major institutions on the treatment of myeloma and leukemia were identified using a methodology similar to that described by Delaney et al.7–12 Currently, there are no published Australian clinical practice guidelines for hematologic malignancies, although they are currently in development. Sources of guidelines used in this study were the comprehensive Physicians Data Query data base of the U.S. National Comprehensive Cancer Institute,13–19 the National Comprehensive Cancer Network,20–22 the British Society of Haematology,23 the British Haematological Society,24 and the European Society of Medical Oncology.25–27 The level of evidence that supported each recommendation for radiotherapy use was classified using the Australian National Health and Medical Research Council hierarchy of levels of evidence.28

The optimal radiotherapy utilization trees (see Figs. 1, 2) were constructed using TREEAGE DATA™ software (version 3.5). Each terminal branch represents “radiotherapy” or “no radiotherapy” as the management decision. Because it was intended to calculate the need for “at least one course of radiotherapy during the illness,” the tree branch was terminated as soon as the first indication for radiotherapy was reached, and potential future reasons to use radiotherapy were ignored.

Figure 1.

Optimal radiotherapy (RT) utilization tree for leukemia. CNS: central nervous system; ALL: acute lymphoblastic leukemia; HLA: human leukocyte antigen; BMT: bone marrow transplantation.

Figure 2.

Optimal radiotherapy (RT) utilization tree for myeloma.

Each branch of the tree signifies an attribute that has an impact on a management decision (e.g., the stage of the tumor). Above each branch in the optimal radiotherapy utilization tree is a description of the specific attribute that has led to the treatment decision. Each number below the branch signifies the proportions of the attribute based on epidemiological data.

Epidemiologic Data

The source with the highest ranking was used to determine the incidence of each radiotherapy indication.28 We used Australian national and state cancer registry epidemiological data wherever available, because the results of this study will be used to plan future radiotherapy facilities in Australia. When national data were unavailable, more specific data sets (such as those of state cancer registries) were used for information pertaining to tumor stage and pathology. Where national or state registry data were unavailable for particular decision-tree branches, other sources were used, such as data bases from other countries or institutional reports. Actual utilization rates were identified and were compared with the optimal utilization data.

Markov Analysis

Because the natural history of myeloma is of an indolent disease with frequent flare-ups over time, a proportion of patients with myeloma who initially are pain free will develop pain later in their treatment course. Although most decision trees include a simple notion of time, this can be modeled better using a state transition model, also called a Markov model.29 Discrete Markov models enumerate a finite set of mutually exclusive possible states such that, in any given time interval, an individual member of the Markov cohort is in only one of these states. A Markov model (Fig. 3) was used to determine the proportion of patients who will require radiotherapy for disease progression, and the data were incorporated into the myeloma tree.

Figure 3.

Markov modeling for myeloma. Note that this branch fits onto the 4 terminal branches entitled “no bone pain” in FIGURE 2. RT: radiotherapy; Init: initial; Incr: increased; Rwd: reward.

The subtree emanating from a state is used to represent the possible transitions from that state: in this case, the proportion of patients each year who develop pain or die. Once they have entered either pain or death, no transitions out of this state are possible (enter an absorbing state). The increment entering each of these absorbing states (pain or death) is calculated from the proportion of patients likely to develop pain each year or die. It was decided to terminate the model after 4 cycles were completed, representing 5 years of disease.29 The subtree (Fig. 3) was inserted at the four terminal branches in which patients had “no pain” after bisphosphonate therapy.

Peer Review

The trees and the epidemiologic data were sent for external review to a multidisciplinary panel of cancer experts (from hematologic, surgical, medical oncology, radiation oncology, palliative care, and nursing backgrounds). The review process was overseen by an independent steering committee of general oncology experts convened by the National Cancer Control Initiative (Australia).

RESULTS

Using the best available evidence, a list of indications for radiotherapy for leukemia and myeloma were created, and optimal radiotherapy utilization trees were generated (Figs. 1, 2). Each terminal branch of the tree showed whether radiotherapy was recommended for patients with those particular attributes.

The optimal utilization rate was calculated by determining the incidence of each indication for radiotherapy. Table 1 shows the proportion of patients with leukemia who should receive radiotherapy based on evidence, and Table 2 shows the incidence of these attributes for radiotherapy in patients with leukaemia. Table 3 shows the clinical situations in which radiotherapy is recommended and the guideline or source of evidence for the recommendation for myeloma, and Table 4 shows the corresponding epidemiological data for myeloma.

Table 1. Leukemia—Indications for Radiotherapy: Levels and Sources of Evidence
Outcome no.aClinical scenarioTreatment indicatedLevel of evidencebReference(s)cProportion of all patients with leukemiad
  • ALL: acute lymphoblastic leukemia; CNS: central nervous system; RT: radiotherapy; HLA: human leukocyte antigen; AML: acute myeloid leukemia; CR: complete response; BMT: bone marrow transplantation.

  • a

    Outcome number corresponds to terminal branch on tree in figure (e.g., Fig. 1-1 refers to Fig. 1, terminal branch 1).

  • b

    Levels of evidence: Level I, systematic review of all relevant randomized studies; Level II, at least 1 properly conducted randomized trial; Level III, well designed controlled trials without randomization (includes trials with “pseudorandomization” or comparative studies); Level IV, case series. National Health and Medical Research Council, 1998. Guide to the development, implementation and evaluation of clinical practice guidelines, Appendix B, p 56.28

  • c

    Numbers refer to the list of references.

  • d

    Total proportion of all patients with leukemia who are recommended for radiotherapy = 0.04 (4%).

Fig.1-1ALL, age < 15 yrs, CNS disease at presentationRTIII13< 0.01
Fig. 1–2ALL, age < 15 yrs, no CNS disease at presentation, high riskRTII130.01
Fig. 1–3ALL, age < 15 yrs, no CNS disease at presentation, not high risk, CNS/testicular recurrenceRTIII13< 0.01
Fig. 1–4ALL, age < 15 yrs, no CNS disease at presentation, not high risk, early recurrence in bone marrow, HLA-compatible donorRTIV13< 0.01
Fig. 1–8ALL, age 15–560 yrs, recurrence, HLA-compatible donorRTIII14< 0.01
Fig. 1–15AML, age 16–54 yrs, low risk, CR to induction chemotherapy, recurrence, HLA-compatible donor, proceed to BMTRTIII15, 21, 400.01
Fig. 1–20AML, age 16–54 yrs, high/intermediate risk, HLA-compatible donor; proceed to BMTRTIII15, 400.02
Table 2. Leukemia: The Incidence of Attributes Used to Define Indications for Radiotherapy
Population or subpopulation of interestAttributeProportion of population with this attributeQuality of informationaReference(s)b
  • ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; CLL: chronic lymphocytic leukemia; CML: chronic myeloid leukemia; CR: complete response; BMT: bone marrow transplantation; HLA: human leukocyte antigen.

  • a

    Quality of information/hierarchy for epidemiologic data: α, Australian National Epidemiological data; β, Australian State Cancer Registry; γ, epidemiological data bases from other large international groups (e.g., Surveillance, Epidemiology, and End Results Program); δ, results from reports of a random sample from a population; ε, comprehensive multiinstitutional data base; ζ, comprehensive single-institutional data base; θ, multiinstitutional reports on selected groups (e.g., multiinstitutional clinical trials); λ, single-institutional reports on selected groups of patients.

  • b

    Numbers refer to the list of references.

All cancersLeukemia0.03α41
All leukemiasALL0.11θ31
 AML0.46  
 CLL0.30  
 CML0.13  
ALLAge < 15 yrs0.45β42
ALL, age < 15 yrsCNS disease at presentation0.02θ43, 44
ALL, age < 15 yrs, no CNS disease at presentationHigh risk0.11–0.13θ44, 45
ALL, age < 15 yrs, no CNS disease at presentation, not high riskRecurrence0.12–0.37θ44, 46–49
ALL, age < 15 yrs, no CNS disease at presentation, not high risk, recurrenceCNS/testicular0.35θ44, 46–49
ALL, age < 15 yrs, no CNS disease at presentation, not high risk, recurrence in bone marrowEarly recurrence0.7ε50
ALL, age < 15 yrs, no CNS disease at presentation, not high risk, recurrence in bone marrowHLA-compatible donor0.35ε45
ALL, age > 15 yrsAge < 60 yrs0.67δ32
ALL, age 15–60 yrsCR0.82ε51
ALL, age 15–60 yrs, CRRecurrence0.5ε52, 53
ALL, age 15–60 yrs, CR, recurrenceHLA donor0.33θ54
AML, age at diagnosisAge < 15 yrs0.03γ32
 Age 16–54 yrs0.28  
 Age > 55 yrs0.69  
AML, age 16–54 yrsLow risk0.23θ55
AML, age 16–54 yrs, low riskCR0.91θ55
AML, age 16–54 yrs, low risk, CRRecurrence0.49θ56
AML, age 16–54 yrs, low risk, CR, recurrenceHLA donor0.33θ54
AML, age 16–54 yrs, low risk, CR, recurrence, HLA-compatible donorRefuse/recurrence pre-BMT0.19θ54
AML, age 16–54 yrs, intermediate/high riskCR to induction therapy0.83θ55
AML, age 16–54 yrs, intermediate/high riskHLA-compatible donor0.33θ54
AML, age 16–54 yrs, intermediate/high risk, HLA-compatible donorRefuse/recurrence pre-BMT0.19θ54
Table 3. Myeloma—Indications for Radiotherapy: Levels and Sources of Evidence
Outcome no.aClinical scenarioTreatment indicatedLevel of evidencebReference(s)cProportion of all patients with myelomad
  • BMT: bone marrow transplantation; RT: radiotherapy.

  • a

    Outcome number corresponds to terminal branch on tree in figure (e.g., Fig. 1-1 refers to Fig. 1, terminal branch 1).

  • b

    Levels of evidence: Level I, systematic review of all relevant randomized studies; Level II, at least 1 properly conducted randomized trial; Level III, well designed controlled trials without randomization (includes trials with “pseudorandomization” or comparative studies); Level IV, case series. National Health and Medical Research Council, 1998. Guide to the development, implementation and evaluation of clinical practice guidelines, Appendix B, p 56.28

  • c

    Numbers refer to the list of references.

  • d

    Total proportion of all patients with myeloma in whom radiotherapy is recommended = 0.38 (38%).

Fig. 2-1Symptomatic, age < 60 yrs, eligible for BMT, recurrence, persistent bone pain after biphosphonates, palliative treatment for bone painRTI19, 22, 240.03
Fig. 2–5Symptomatic, age < 60 yrs, unable to tolerate initial therapy, palliative treatment for bone painRTI19, 22, 240.02
Fig. 2–7Symptomatic, age < 60 yrs, ineligible for BMT, recurrence after initial therapy, palliative treatment for bone painRTI19, 22, 240.01
Fig. 2–9Age > 60 yrs, persistent bone pain after biphosphonates, palliative treatment for bone painRTI19, 22, 240.32
Table 4. Myeloma: The Incidence of Attributes Used to Define Indications for Radiotherapy
Population or subpopulation of interestAttributeProportion of population with this attributeQuality of informationaReference(s)b
  • BMT: bone marrow transplantation.

  • a

    Quality of information/hierarchy for epidemiologic data: α, Australian National Epidemiological data; β, Australian State Cancer Registry; γ, epidemiological data bases from other large international groups (e.g., Surveillance, Epidemiology, and End Results Program); δ, results from reports of a random sample from a population; ε, comprehensive multiinstitutional data base; ζ, comprehensive single-institutional data base; θ, multiinstitutional reports on selected groups (e.g., multiinstitutional clinical trials); λ, single-institutional reports on selected groups of patients.

  • b

    Numbers refer to the list of references.

All registry cancersMyeloma0.01α41
MyelomaSymptomatic0.97λ57
Myeloma, symptomaticAge < 60 yrs0.22γ32
Myeloma, symptomatic, age < 60 yrsSuitable for BMT0.86ε58
Myeloma, symptomatic, age < 60 yrs, suitable for BMTAble to complete BMT0.78ε58
Myeloma, symptomatic, age < 60 yrs, suitable for BMT, able to complete BMTProportion with recurrence0.62θ59
Myeloma, symptomatic, age < 60 yrs, suitable for BMT, able to complete BMT, recurrenceSuitable for salvage treatment0.89θ59
Myeloma, symptomatic, age < 60 yrs, suitable for BMT, able to complete BMT, recurrence, suitable for salvage treatmentBone pain after bisphosphonates0.42θ60
Myeloma, symptomatic, age < 60 yrs, suitable for BMT, unable to complete BMTBone pain after bisphosphonates0.42θ60
Myeloma, symptomatic, age > 60 yrsBone pain after bisphosphonates0.42θ60

The outcome numbers in Tables 1 and 3 correspond to the outcome positions in the tree. The last (far right) column represents the incidence of each clinical indication for radiotherapy as a proportion of patients diagnosed with that type of cancer. Tables 2 and 4 show the epidemiologic data corresponding to each branch point and the source of the data as well as the hierarchical level of the data obtained.

Outcomes

Leukemia

There were 25 possible “outcomes” for this tree, and 7 of those outcomes recommended that radiotherapy be considered (5 outcomes for acute lymphoblastic leukemia and 2 outcomes for acute myeloid leukemia). The optimal radiotherapy utilization rate for leukemia was calculated t 0.04, i.e., 4% of all patients with leukemia in Australia should receive radiotherapy at some point during their treatment based on the best available evidence. The proportions of patients with acute lymphoblastic leukemia and acute myeloid leukemia for whom radiotherapy is recommended were 15% and 6%, respectively.

Current practice

Data on the radiotherapy utilization rate for leukemia were identified from three major sources: South Australia; the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) Program; and Northern England.30–32 The actual utilization rate for radiotherapy for leukemia was 3–6% (Table 5). The only actual Australian radiotherapy utilization data were from the South Australian Cancer Registry, which reports on utilization rates for the population of leukemia patients who are treated as part of their primary course of care. That registry reported report that 6% of patients diagnosed with leukemia were treated with radiotherapy.

Table 5. Comparison of Optimal and Actual Radiotherapy Utilization Rates for Myeloma and Leukemia
DiagnosisOptimal RT utilization rate (%)Actual RT utilization rate (%)
South AustraliaaSEERbNorthern and YorkshirecMayo Clinic, 1975d
  • SEER: Surveillance, Epidemiology, and End Results Program.

  • a

    See South Australian Cancer Registry, 2000,31 and Luke et al, 2004.37

  • b

    See U.S. Department of Health and Human Services, 2000.32

  • c

    See Northern and Yorkshire Cancer Registry, 1999.30

  • d

    See Kyle, 1975.38

Myeloma3834252455
Age < 55 yrs 373330 
Age 55–74 yrs 272828 
Age > 74 yrs 191818 
Leukemia4653 

Myeloma

There were 11 possible outcomes for myeloma, 4 of which recommended radiotherapy. All of these outcomes were for palliative treatment of bone pain, which is the most common indication by far for radiotherapy for myeloma. The optimal radiotherapy utilization rate for myeloma is calculated at 0.38 (i.e., 38% of all patients with myeloma in Australia should receive radiotherapy at some point during their treatment based on the best available evidence).

Markov Modeling

To assess the need for radiotherapy at later stages of myeloma, which has a chronic, recurring nature, we used a Markov model. This analysis required data on the proportion of new events per year and the death rate, which were identified from two studies that were established to examine the efficacy of bisphophonates in the management of myeloma. Berenson et al. reported on the number of skeletal-related events that occurred among myeloma study patients over a 2-year period while on treatment with bisphosphonates. The increase in skeletal events, between 9 months and 21 months, was 17% and 19% when no pamidronate or pamidronate was administered, respectively.33, 34 These proportions are higher than the actual numbers of patients who received radiotherapy reported in this article, in which there was an increase of 22% reported for the pamidronate arm and 12% reported for the nonpamidronate arm. The lowest proportion of 12% was used in the Markov model. The proportion of deaths was calculated from graphs displayed in the article by McCloskey et al.35 Within the first year, approximately 20% of patients died (read directly from the survival graph, because this was not reported), which subsequently fell in the second and third years to 16% and 14%, respectively.35, 36

To estimate the proportion of new events occurring annually that require radiotherapy, the lowest increase in proportions for bone pain (12% per annum) and the highest annual death rate (20%) were used to model the development of new bone pain that would require radiotherapy in the Markov model. The reason for choosing the highest death rate and the lowest bone pain incidence rate was to prevent an overestimate of the radiotherapy utilization. The cycle was repeated 4 times to give an overall proportion for 5 years. The proportion of new patients requiring radiotherapy was 26% for episodes of bone pain over 5 years. This proportion was added to the original tree to model the actual radiotherapy utilization rate and increased the overall proportion that required radiotherapy from 38% to 64%.

Current Practice

There were data available on the actual radiotherapy utilization rates from several sources. These included South Australia,37 SEER,32 and Northern England.30 The actual utilization rate for radiotherapy for myeloma ranged from 24% to 34% (Table 5). The actual utilization rate for radiotherapy for leukemia was 3–6%.

DISCUSSION

In the current study, we report that 38% of patients with myeloma should receive treatment with radiotherapy at some time during their illness according to available evidence and epidemiological data. When the Markov model was included, this increased to 64%. The only actual Australian radiotherapy utilization data were from the South Australian Cancer Registry, which reports on utilization rates for the population treated as part of their primary course of care within 12 months from diagnosis. That registry reported that, among patients who were diagnosed with myeloma, 34% were treated with radiotherapy.

SEER and Yorkshire reported lower rates compared with South Australia, with 23–25% of patients diagnosed with myeloma receiving treatment with radiotherapy. A greater proportion of younger patients than older patients were treated with radiotherapy.30, 32

A population-based study reported in 1975 that 55% of patients diagnosed with myeloma were treated with radiotherapy.38 This may reflect the need for radiotherapy over a longer period than the other reports, and it also predates the use of bisphophonates. The randomized trial by Berenson et al. included radiotherapy utilization, and those authors reported that the actual proportions of patients requiring radiotherapy by 21 months were 50% and 34% in the nonpamidronate arm and the pamidronate arms, respectively.34

Actual radiotherapy treatment rates are reported for specific periods of time after diagnosis, usually as initial treatment at diagnosis or within a defined period rather than treatment at anytime during the illness. Treatment initiation for symptomatic disease progression, recurrent disease, or retreatment rates are difficult to factor into actual rates. Other differences include the interpretation of guidelines and different guideline indications for radiotherapy among countries and centers and the personal beliefs of referring clinicians.

The recommended proportion of patients in Australia who, according to the best available evidence, should receive at least a single course of radiotherapy for leukemia is 4%, which corresponds with actual practice. The actual utilization rate for radiotherapy for leukemia was 3–6%. It is surprising that actual and optimal utilization rates correspond, because it has been found that the more common malignancies have substantially lower utilization rates compared with evidence-based estimates,39 although this may reflect the fact that children with leukemia and bone marrow transplantation recipients usually are treated on protocols and managed in tertiary referral centers in contrast to more common malignancies, which are treated across the community. The optimal radiotherapy utilization for other hematologic malignancies61 and other cancers7–12 are also being developed.

Acknowledgements

The authors thank the reviewers: Professor John Seymour, Professor Ken Bradstock, Professor Alan Rodgers, Dr. Katy Clark, Dr. Peter O'Brien, Dr. Gary Pratt, Dr. David Christie, Dr. Robert Lindeman, and Dr. David Speakman and the members of the steering committee of Australian National Cancer Control Initiative for helpful comments regarding the study design and optimal radiotherapy utilization trees.

Ancillary