The clinical value of routine whole-body magnetic resonance imaging (MRI) in palliative care

MRI has throughout the last decade become a more available and faster imaging method. This has increased the use as well as given new possibilities of using MRI. Today it is fully possible, within a reasonable amount of time, to perform a whole body MRI, even in fragile patients (1–3). In several institutions this is a procedure in use for the detection of bone metastases (4). Whole-body MRI (WBMRI) is thus a frequently used method, which might easily be introduced for different investigative purposes (5–7).

Patients referred for palliative treatment are fragile and have a short life expectancy. In such a setting it is questionable to include an extensive examination such as MRI. The main goal is to reduce the symptom burden (8, 9). MRI has the possibility to detect lesions that cause or are likely to cause pain. Thus, palliative care might be planned better ahead and the treatment more efficiently performed (10). The findings might be bone metastasis or large lesions causing pressure on vital or painful structures. Based on this information it might be possible to give palliative radiation therapy, more focused drug therapy, or even minimal invasive therapy in order to relieve pain.

A hospitalized patient undergoes a series of procedures for diagnosing and treatment during a hospital stay, or during the entire treatment of a disease. It is difficult to find a final outcome for the patient, and if so, this outcome cannot tell anything about the outcome of a single procedure during a long hospital stay, or an entire treatment, with several procedures. It is, nevertheless, important to find outcomes or values of different procedures (11, 12). The clinical usefulness may be such an endpoint in radiology. Accordingly, in radiology the value could be an objective evaluation of the clinical usefulness of an investigated procedure. This is what we have chosen to call clinical value or clinical utility in this study. However, different definitions exist (13–15).

The aim of the study was to evaluate the value, or clinical utility, of introducing an advanced imaging procedure in order to improve palliative care. We intended to do this by evaluating the clinical utility of any possible new information gained, and discuss it in view of the use of resources and possible ethical questions.

Material and Methods

All patients admitted to the Department of Palliative Care, from June 30, 2006 to January 2, 2008 were eligible for inclusion. In total 20 patients were included (of 190 patients admitted to the department). The initial inclusion criteria were to include all patients referred to the department of palliative care. Exclusion criteria were: patients not compatible with MRI (pacemakers, claustrophobia, certain metal implants, etc.); patients who did not wish to undergo further examinations; patients who were not considered appropriate for further investigation due to their medical status; or their treating physician considered it inappropriate to include them. In total 170 patients were excluded. The reasons for exclusion were: terminal patients who died during the stay = 98 (excluded by their physicians); short visit and moved to nursing home = 20; no pain-related problems = 16; other problems than pain-dominated = 13; no diagnostic problems; unethical to include = 8; declined participation = 7; transferred to another department = 5; not asked = 2; not a cancer patient = 1.

The included patients were referred consecutively for MRI as they were admitted to the hospital. The result of the MRI was submitted to the referring physician who registered prospectively whether or not the radiology report added new information, had any impact on treatment or led to a change of treatment or introduced new treatment. The registration was done with a standardized questionnaire (Appendix).

A panel of physicians from the Department of Palliative Care retrospectively evaluated the clinical records in order to see if the increased information had any impact on the total handling of the patients (clinical utility). The patients were evaluated at endpoint = death.

The MRI examination was performed on a GE-Signa 1.5 T from 2002, with 33 mT/m Gradient (GE Healthcare, Milwaukee, WI, USA). The images were obtained in a coronal view only. The coronal view was chosen in order to achieve the best overall view of the body, despite that it might have been better to use sagittal views of the spine. T1W and T2W STIR sequences were used. In order to reduce stress to the patients, a rapid protocol was introduced, partly at the cost of optimal images. During the study, we had an update of the MRI and the present protocol is: Coronal 2D sequence: TE 27, TR 4960, Inv time 120, FOV 48, slice th 10, spacing 0, NEX 2. 3D coronal sequence: TE In phase 4.2 → 14.0, flip angle 25, FOV 48, slice th 5, NEX 1. Resolution was 256 × 256 for the present 3D sequence, the 2D sequences has a resolution of 512 × 256.

All patients had a cancer diagnosis before being referred to palliative care. Primary diagnoses were: colorectal 8, urogenital (including the prostate) 6, pancreas 3, lungs/pleura 3. The group included 14 male and six female patients. At the examination time the average age was 64 years (median age 62 years, range 40–89 years). The average time to death after the MRI examination was 88 days (range 4–592 days), and the median time 30 days.

The possible distress of undergoing an MRI and possible meagre outcome raises ethical questions that we had to consider. Even though we reduced the MRI protocol as much as possible, the majority of the submitted patients had to be excluded. Combined with experience of MRI in earlier patients, we considered the included patients to have a gain in treatment, with an acceptable amount of distress.

The Regional Ethics Committee approved the investigation. All patients received oral and written information, and signed a written letter of consent. In order not to cause any distress to these patients, both the patients themselves, and the physician responsible for the patient, could at any time withdraw from the study.

Results

Sixteen patients (80%) had new findings that could be of importance to the patient (Table 1). In 13 patients these findings could theoretically influence the treatment (Table 2), and for 40% (8 patients) there was a change in treatment due to the MRI result. Two patients had metastases not previously recognized (Tables 1 and 2).

OPEN IN VIEWER

Table 1

Prospective registration of MRI results, and their consequences

	Metastases		Important		Change of treatment
No.	Yes	No	Yes	No	Yes	No
1	X		X			X
2	X		X		X
3	X		X			X
4	X		X			X
5	X		X			X
6		X	X		X
7		X	X			X
8	X		X			X
9	X		X			X
10	X		X			X
11	X		X		X
12	X		X		X
13	X		X			X
14	X		X			X
15	X		X			X
16	X		X		X
17	X		X		X
18	X		X		X
19	X		X			X
20	X		X		X
Sum	18 (90%)	2 (10%)	20		8 (40%)	12 (60%)

These results are based on a prospective registration done by the physician at the department of palliative care. Three columns: registration of metastases, evaluation of importance performed by the patients' physicians, and objectively registered change in treatment

OPEN IN VIEWER

Table 2

Probable consequence without the use of WBMRI, a retrospective evaluation

	To treatment		To the patient
No.	None	Delay	No treatm.	None	Small	Large
1	X				X
2		X				X
3		X			X
4	X			X
5	X			X
6		X				X
7	X			X
8	X			X
9			X			X
10			X		X
11		X				X
12		X				X
13	X				X
14		X			X
15	X				X
16		X				X
17		X				X
18		X				X
19		X			X
20		X				X
Sum	7 (35%)	11 (55%)	2 (10%)	4 (20%)	7 (35%)	9 (45%)

This table presents a subjective evaluation done by the clinicians, in order to establish consequences and thus the value of introducing WBMRI. It must be read in connection with the last part of Table 1. It is a comparison with a zero-alternative. The clinicians evaluated retrospectively and estimated what the result would have been without any imaging procedure to reveal the results shown by MRI. The first part is an evaluation of the consequences for the treatment without MRI: no consequences, a delay of treatment or no treatment. The latter is an evaluation of the consequences for the patients without MRI: No consequences, small consequences or large consequences. The most important result here is the evaluation saying that to nine patients it would have been of major consequences not to have MRI. Patient number 9 had a negative MRI result. This was of great importance, but did not lead to a change in treatment, corresponding to the eight patients with change of treatment (Tables 1 and 3)

The detailed results of the eight patients, who had a change of treatment directly after the MRI examination, are presented in Table 3.

OPEN IN VIEWER

Table 3

Change in treatment

No.	Earlier radiology	MRI findings	Change in treatment
1	CT + colon, colonic cancer	Widespread skeletal mets	Change in palliative treatment
2	CT, large tumor	Pelvic abscess in tumor area	Given antibiotics, prolonged life + pain relief
3	CT + X-ray, seen skeletal mets	Compression of the spinal cord	Radiation therapy
4	CT, skel mets	Widespread disease	Change of palliative treatment radiation impossible
5	CT	Skeletal mets growing into the gluteus muscles	Radiation therapy
6	US + X-ray	Pleural effusion + skel mets	Pleurocentesis
7	CT	Large tumor masses in the thorax and thoracic wall	Radiation therapy
8	X-ray	Skeletal mets	Radiation therapy

This is a presentation of the eight patients who did receive a change of treatment or a direct intervention as a result of the MRI examination. The radiology examinations mentioned as earlier examinations are those that are within a reasonable time limit

Radiology diagnosis mentions the part of the diagnosis that resulted in a change of treatment

The last column states what the direct intervention was

CT = computed tomography; Mets = metastases; X-ray = plain chest X-ray

Fig. 1 shows an example of bone metastasis, the most frequent findings in these patients.

OPEN IN VIEWER

Fig. 1

All the vertebrae have a pathological low signal due to widespread metastases. Compare the spinal vertebrae to the shoulders

Table 1 presents the consequences/gain of the information. A delay in treatment to these patients may result in more pain during the last part of their lives. Tables 2 and 3 present in more detail the change induced by the MRI result, as evaluated by the clinicians. Four had radiation therapy and added analgesics, not planned to be instigated before the MRI result was known, one a pleurocentesis and added analgesics and three changes in medication (more opioids and steroids).

Retrospectively the clinicians did a subjective evaluation whether or not the results of the MRI examination did have significance. They answered the following questions as yes/no:

Significance for further medical investigation of the patient 12/8;

Importance to the patients' interests 16/4;

Significance to the ongoing treatment 11/9.

All the patients had performed other radiological examinations during their disease, 17 less than a month ago. Six of the eight patients who had their treatment altered had conventional X-ray, computed tomography, or ultrasound less than a month before the MRI. Of the 12 patients who did not get a change in treatment, four had a radiological examination less than a month ago.

To what degree the radiology results of previous examinations were altered by the WBMRI is presented in Table 3.

Discussion

The main findings of the study were that a WBMRI might detect changes that could influence the treatment in a majority of patients admitted to a palliative department, and actually caused a change in treatment for nearly half of the examined patients.

The number of 20 patients was small compared to the total number of admittances to the department. This was unavoidable, as we were concerned with not adding any pain or distress to these patients. The referring physicians did not consider more than 20 patients suitable (more ‘pain than gain’). This is a common challenge in palliative care, as the patients are fragile and it would be unethical to submit patients to potential stressful procedures. We know that this group of patients would include patients at the end stage of their lives, with already heavy medication, as well as patients who did not wish to have more examinations or therapy attempts. The exclusion criteria had to be accordingly wide. No attempt at all to persuade patients to participate was made, nor did the patients' physicians refer patients if they intended not to change the ongoing treatment. All this must be anticipated to give a bias.

The number of attending physicians is small in such a department. The first registration the handling physician registered without any discussions with colleagues or subjective comments. It ought thus not to be biased, but is still only the opinion of one physician. The consensus might be biased due to the close working relationship between a small number of physicians.

However, our results indicate a reasonably robust conclusion. We have found progress of disease and new foci of disease in a majority of the patients. This was not surprising as the patients had severe cancer diseases. It is, nevertheless, remarkable that the clinicians could so often improve treatment due to the examination. Perhaps not surprising, the detection of bone metastases was the most significant. Radiology is frequently being used for this purpose (16).

With such a small number of patients, statistics become dubious, and we have found it better to present all data for each patient. Some of the data indicating a significant value of the MRI's usefulness are so robust that they would also be statistically significant. When scrutinized in detail, we see that eight of the patients did objectively have a change in treatment.

We had two main questions to answer: one was to find progress of disease that could be better handled if recognized and the second was registering any changes due to radiologic findings. Again, the results are robust enough to at least indicate that we have a reasonable conclusion. The effect in the hospital of this investigation is that the physicians at the department of palliative care use radiology more often for better planning of treatment. This is a sort of clinical acceptance of the clinical usefulness of the procedure. They do, however, use WBMRI or more dedicated MRI procedures according to clinical questions and needs.

The impression of the work as it went on was that we ought to use advanced methods such as MRI more often, in order to help clinicians add welfare to the patients. This study was not performed in order to improve WBMRI. Our aim was to see how it could be used in order to improve treatment of a small and vulnerable group of patients. The study does not say, but might imply that other methods, such as whole-body CT and CT-PET may be beneficiary as well. For practical reasons we could choose either CT or MRI as method for this group of patients. We chose MRI rather than CT because we expected bone metastases to be an important finding (17).

In daily practice we often question whether or not a patient with a short life expectancy should have advanced imaging. ‘Isn't it a waste of resources?’ one would ask. By doing this, we probably refuse optimum treatment to a relatively large number of patients. It inevitably raises an ethical question: What is the optimum use of resources? The economists would probably refer to a Pareto principle, even with all its inadequacies (18), but to doctors it raises ethical considerations on how to use resources. Another question is to what extent we are legally or morally obligated to support the patients with the optimal diagnosis and treatment? Nobody has discussed these problems exactly, but similar questions have been discussed (19).

We have demonstrated that a group of patients needing palliative care to relieve suffering would benefit from more use of MRI. By this we have also indicated that the use of ‘high-tech’ imaging might be beneficiary to other groups of patients. It is also a way of adding value to the patient, discussed in connection with making health services more efficient, even though what we present is a new way of thinking (20). These are thoughts and questions for debate. They challenge how we prioritize, how we deal with weak groups of patients, and they challenge our ethical views concerning the use of limited medical resources. What is the value of relieving pain? Is it ethical or unethical to limit resources to patients with a short life expectancy? We have not found articles or books handling this ethical dilemma. Discussing principles of ethics concentrate on the patients' autonomy or principles in general when handling ethical issues (21, 22). One book did discuss QUALYs; life expectancy and prioritizing resource-based life expectancy and old versus young (23). However, how can we deny a person in pain the optimal pain relief? The 15–20 minutes used is not a major use of resources, and could therefore be justified. The difficult question is where to put limits: MRI in order to examine possible causes of pain, of cancer or other conditions might be acceptable, but introducing CT-PET might be too much.

This study has had some influence on our clinical routines. WBMRI is more in use, but it has not become a screening method. However, the general use of MRI has increased, and is demanded according to pain or other clinical symptoms, but often more specified whether it is WBMRI, skeletal MRI, MRCP, or other procedures.

In conclusion, we have demonstrated that palliative patients could benefit from a relatively short MRI examination. We would advocate that patients with severe pain caused by advanced disease should have the optimum treatment for their pain, and that might include MRI. We believe that the use of resources is small compared to the benefits gained.

Conflict of interest: None.