MRS gains indications, but still fights for reimbursement

To paraphrase Mark Twain, reports of the death of diagnostic MR spectroscopy are greatly exaggerated. CPT 76390 is considered standard of care as an effective imaging technique for the diagnosis, treatment, and monitoring of patients with brain lesions by Cigna Healthcare, a respected healthcare provider,¹ though declared "investigational" by Blue Shield, Anthem, and Medicare. Radiologists and other physicians are confused and annoyed by some insurers' refusal to reimburse for their MRS services.

MRS as a diagnostic modality is far from death. It is actively and generously reimbursed by a majority of healthcare plans and private insurers. But MRS is the focus of a healthy tussle between two medical disciplines: radiological and medical science on the one side, and evidence-based medicine (EBM) on the other.

MRS proves its diagnostic worth every day in the brain, where 10 or more urgent diagnostic challenges are recognized (Table 1), the prostate, and the breast, in which it affects cancer diagnosis and patient outcomes. As for EBM, a computerized search of the literature will turn up more than 20,000 hits for diagnostic MRS.

Yet diagnostic imaging has come to different conclusions from those of organized EBM. We can fix this breach, so that our patients and clinical practices can employ the latest techniques with reimbursement. A brief reiteration of the arguments for and against MRS suggests a game plan to effectively preserve its use and expand the present limited CPT code to a solid program.

Clinical MR spectroscopy is a simple, noninvasive add-on sequence that dramatically enhances the value of routine brain MR imaging, and it is emerging as a valuable technique in many organs and tissues. It is in the brain, however, that the real diagnostic value lies. For radiologists, the fact that MRS produces a spectrum rather than an image has long been a barrier to performing, reading, and reporting.

Figure 1 shows that this is a needless worry. A brain spectrum is an image like any other, with a series of peaks that can be read from right to left. These identify seven brain metabolites: lipid, lactate, N-acetylaspartate (NAA), glutamate and glutamine, creatine, choline, and myo-inositol. The popular mnemonic for remembering the metabolites in order is Lying Lazy No-Good Crooks Chollected My Insurance. The height of the peak is its concentration, usually measured as a ratio to the height of the central creatine peak. The steep angle from left to right, here manifest as the climbing angle of a jet aircraft (Hunter's angle), identifies the normal relative peak ratios for the adult brain.

FOCAL BRAIN DISEASES

Nothing could be simpler than diagnosing brain tumor with MRS, as shown in Figure 2, where the Hunter's angle is clearly reversed.² The spectrum shown was acquired from the voxel indicated in the right posterior parietal lobe, where neuronal marker NAA is much reduced and the membrane marker Cho is markedly increased. Ancillary findings in the spectrum, such as the presence of unusual excess of lipid and lactate, add detail to the interpretation. Abscesses, for example, will often contain additional metabolites such as succinate and alanine, providing further differential details in the spectrum (Figure 3).

NONFOCAL BRAIN DISEASES

Although tumor diagnosis is one of the most described uses of MRS, it is also helpful in assessing other conditions involving the brain.

- Dementia, including Alzheimer's disease. MRS represents a useful examination in patients with nonspecific changes or without focal changes on MRI. Figure 4 illustrates the remarkable and diagnostic features of a brain spectrum acquired from gray matter in the posterior cingulate gyrus (not hippocampus) in a patient with clinically verified Alzheimer's disease. NAA/Cr is reduced, the result of neuronal loss, while mI/Cr is increased, without significant abnormality in Cho/Cr. The latter is unexplained but appears very specific, since it is observed in patients with mild cognitive impairment (MCI, not shown), a precursor of Alzheimer's disease increasingly recognized in the aging U.S. population. As shown in Figure 4, familial AD patients, Down syndrome adults with dementia, and even children with Down syndrome show some or all of the same MRS features.

- Hypoxic encephalopathy, including persistent vegetative state post-cardiac arrest. MRS peak height is a quantitative measure of neuronal density and therefore offers very precise prognostic information after global hypoxic brain injury. This information can be obtained by performing MRS on the comatose patient when MRI is normal and neurological outcome is completely obscure to clinical testing. Progressive degrees of reduction in NAA, the appearance of lactate, lipid, and Glx, as well as the timing of MRS in relation to the insult, all add precision to this important decision.

- Stroke. Localized MRS can be used to determine the state of neuronal health and predict future recovery. Figure 5 indicates the ideal, in which anatomic and functional, as well as biochemical, data contribute to the diagnosis and management of stroke. Note the huge excess of lactate and bright diffusion-weighted imaging within the distribution of the occluded right middle cerebral artery. It has long been expected that MRS applied in stroke would add precision to the definition of penumbra and hence assist in active therapy. Prospective studies on therapeutic impact that are clearly needed have been delayed unnecessarily by the concern that MRS takes too long in an already packed imaging protocol. In the example shown, chemical shift imaging took 2.4 minutes and the entire examination no more than 25 minutes.

- Normal brain development and neonatal hypoxic-ischemic encephalopathy. Dramatic changes in brain MRS occur from birth (Figure 6), reflecting normal brain development and permitting precise "dating" of the normal newborn. Seen in terms of Hunter's angle-reversed at birth, horizontal at two months, and near the adult angle by two years-age-related normal curves are used by the radiologist when reading and interpreting the newborn spectrum. Examples are shown in the figure. There is no difficulty in determining neurological outcome after putative perinatal hypoxia.

- Closed-head injury. Head injury is a good example of combinations of focal and generalized brain disorder. MRS is of little additional value when applied to the region of dramatic focal abnormality but extraordinarily useful in defining the state of the remaining, seemingly normal, brain. This can be the subject of diffuse axonal injury or hypoxic injury. An inverse relationship between NAA/Cr and late clinical outcome has been established in several studies.

- Epilepsy. Epilepsy is often the result of metabolic brain disease, and such patients can be identified by MRS. More frequently, MRS has been employed to lateralize mesial temporal sclerosis before ablative surgery.

- Metabolic brain disorders. MRS, a metabolic probe, is obviously valuable for the differential diagnosis of children with inborn errors of metabolism and neurological symptoms. Less well recognized is the concept of MRS use for the elucidation of systemic diseases that have neurological manifestations of coma, mania, confusion, or tremor. It is advantageous for the radiologist (and the neurologist) to look further than routine brain MRI in such situations.

DIAGNOSES OUTSIDE THE BRAIN

Examinations of breast masses in women (Figure 7) and the prostate gland in men (Figure 8) are perhaps the best developed clinical techniques for the radiologist. Although each is far from routine, since both currently require specialized radiofrequency coils and research software, the central message is the same as that learned from neurospectroscopy: Elevated choline represents advancing tumor, and its disappearance represents therapeutic success.

EVIDENCE-BASED MEDICINE

MRS first received FDA approval in 1992, followed by automated MRS sequences (PROBE)³ in 1995. This noninvasive diagnostic tool assessing brain metabolite concentrations is widely available on any current clinical MR system. In 1998, Dr. Victor Haughton and his colleagues applied and received a CPT 4 code necessary for the billing of MRS: CPT 76390.⁴ Unfortunately, FDA approval and CPT 4 coding were not all that was necessary to convince medical insurance companies that MRS was a clinical tool and not investigational.

From 1998 to 2000, individual healthcare providers, such as Huntington Medical Research Institutes, fought the battle against this categorization, slowly convincing private insurers and healthcare management organizations that MRS was medically necessary and noninvestigational. It was during this time that Huntington worked closely with third-party billers to develop billing strategies unique to MRS that ensured high rates of reimbursement. The number of private insurers paying for MRS increased 30-fold, with 90% reimbursing at the full fee rate. Medicare payments, which did not receive any reimbursement in 1998, increased to 60% of exams being paid at full Medicare rates. A review of Health Care Financing Administration coding for CPT 76390 revealed that from Nov. 1, 1999, to Nov. 1, 2001, CPT 76390 was an acceptable noninvestigational code that received reimbursement.⁵ At this time, reimbursement for spectroscopy was at an all-time high.

In 2001, several issues arose that suddenly placed CPT 76390 in the spotlight. Anthem (Blue Cross Blue Shield of Virginia) ruled that MRS was investigational based on an antiquated study from 1996 with few references from the literature current at that time. The American College of Radiology responded to the ruling but received no response from Anthem. At that time, the ACR also noticed that CPT 76390 was being reimbursed irregularly across different Medicare state organizations. In December 2001, the ACR made a formal request for national coverage of spectroscopy to the Centers for Medicare and Medicaid Services.⁶ During the same month, at the GE MR Masters Series course, rumors spread that CPT 76390 had been removed as a CPT code. A closer examination of that issue revealed that CMS had merely removed MRS from the self-referral category, implying that physicians cannot self-refer themselves MRS exams. This likely stemmed from abuse of CPT 76390 but was not documented by CMS.

With all of the attention drawn to MRS, the CMS response was to do exactly the opposite of what the ACR had requested. In April 2002, CMS withdrew coverage for CPT 76390 altogether.⁷ At the same time, Kaiser Permanente issued an internal review of MRS to determine whether it could be used to differentiate radiation necrosis from recurrent brain tumor. In response to the CMS denial of coverage, the ACR made a formal request for CMS to review CPT 76390 specifically in the context of brain tumor diagnosis in September 2002.⁶ This was done in the hopes that if at least one indication was covered, other indications could subsequently be argued later. Again, with the assistance of Huntington, the ACR and the American Society of Neuroradiologists released practice guidelines for MRS in January 2003.⁸

The Agency for Healthcare and Research Quality (AHRQ) Technology Assessment by Jordon et al⁹ proved to be disastrous to the ACR strategy. This study, commissioned by both the ACR and CMS, concluded that spectroscopy was still an investigational tool according to EBM criteria. Two months later, the Kaiser Permanente internal review made public in conjunction with a Blue Cross Blue Shield technology assessment by Mark et al arrived at the same conclusion.¹⁰ This resulted in nationwide memos released to all Medicare providers and BC/BS providers that MRS was investigational and would no longer be covered after January 2004.¹¹

Using the Fryback and Thornbury criteria (1. technical feasibility, 2. diagnostic accuracy, 3. diagnostic impact, 4. therapeutic impact, 5. outcome impact, 6. societal impact), we surveyed over 28,000 papers and published abstracts identified in PubMed by the keywords "MRS, proton, clinical, diagnosis." We ran the literature search four times to ensure reproducibility and analyzed each printout using random-number tables to ensure relevance. To examine and fairly categorize the extensive publications, we employed the criteria advocated by Fryback and Thornbury and designed an "FT index." The FT index normalizes objective technical diagnosis compared with subjective "outcome" variables by which criteria 1 and 2 are denominators of criteria 3 to 6.

We found that MRS fulfills FT 1 and 2 in more than three independent studies for 13 common brain diseases or syndromes (Table 1). The methodology of the studies had no impact on the technology assessment criteria 1 and 2, which were the most frequently encountered barriers for the nonspecialist, resulting in unjustified exclusion from further analysis in all published EBM studies. When applying the FT index to the Jordan AHRQ study, where only one study qualified for criteria 3 to 6, the FT index is near zero. However, when applying the FT index based on the criteria described above, we found in four separate analyses that the FT index ranged from 0.5 to 1.2, with an average of 0.9. This implies that almost 50% of the papers addressed criteria 3 to 6. Although this study is not in depth, these results indicate that EBM should be researched and analyzed with greater input from those familiar with spectroscopy.

SHORT-TERM STRATEGY FOR REIMBURSEMENT

The Huntington Medical Research Institutes' clinical practice expected to suffer in 2004 from the removal of reimbursement by Medicare. With aggressive billing strategies based on previous experience, however, not only did the practice not suffer a loss, but in fact it saw a gain of 10% from the previous year. Several reasons explain this contradictory observation.

First, Huntington has developed unique billing strategies for MRS that have been detailed in the GE MR Masters Series course offered every year. They include simple measures such as reporting spectroscopy in a separate form from MRI so that HMOs can readily see the added value of the exam. A detailed section of the report demonstrates why MRS was medically necessary for the patient exam. In conjunction with such practice changes, the third-party biller is in a much better position to argue for medical necessity, which would also include statements of medical necessity from the referring physicians as well as aggressive rebukes of denial of payments. This has resulted in a highly successful reimbursement rate of 80% to 90% of cases examined at Huntington.

By using the value-added service of spectroscopy, which can readily delineate Alzheimer's disease from depression and other forms of dementia by using an objective chemical measurement, referring physicians are more likely to send their patients for a combined MRI and MRS exam. The MRI exam, which is fully reimbursable, may take 25 minutes of the total exam time, whereas MRS can make a diagnosis of AD within five minutes. Yet MRS is reimbursed at roughly the same rate as MRI. This implies that even if 83% of the patients examined in a combined MRI/MRS exam were not reimbursed, the remaining 16% that are reimbursed can make up for the small cost expenditure of additional time. Needless to say, much more than 16% of the patients examined at Huntington are reimbursed, such that the cost benefits of CPT 76390 to the radiologist are immediately evident.

LONG-TERM STRATEGY

MRS has taken years to achieve its present prominence for effective clinical use as an indispensable diagnostic tool in the brain, prostate, and breast. These strategies are only temporary solutions to counter the growing perception that MRS is not clinically useful. It is critical that more studies qualify under the criteria established by EBM. Assuming that all agree on technology standards, the common factors for what would constitute an approved study from the EBM perspective are shown in Table 2.

Scientists and clinicians as well as journal reviewers and editors must begin to focus on the last four criteria. Involvement of referring and treating physicians in this process is long overdue. Sales of MRI equipment to radiologists should not blind manufacturers to the nature of their true consumers: HMOs, private insurers, and ultimately the government. If these issues are not addressed in the near future, it is inevitable that spectroscopy will become investigational and reimbursement as we know it will disappear forever.

References

1. Magnetic Resonance Spectroscopy (0244). Cigna HealthCare Coverage Position 2004 11/15/2004 [cited 2005 March 25]; Available from: http://www.cigna.com/health/provider/medical/procedural/coverage_positions/medical/mm_0244_coveragepositioncriteria_magnetic_resonance_spectroscopy.pdf.

2. Lin A, Bluml S, Mamelak AN. Efficacy of proton magnetic resonance spectroscopy in clinical decision making for patients with suspected malignant brain tumors. J Neurooncol 1999;45(1):69-81.

3. Webb PG, et al. Automated single-voxel proton MRS: technical development and multisite verification. Magn Reson Med 1994;31(4):365-373.

4. Prost RW, et al. Recent advances in MR spectroscopy expand its applications in neurologic disease. WMJ 1997;96(11):41-45.

5. Federal Register Archives, Office of the Federal Register, ed. 1999-2001, National Archives and Records Administration.

6. Kaiser CP. CMS studies MR spectroscopy reimbursement. Diagnostic Imaging Online 2002.

7. Medicare Program. Revisions to payment policies and five-year review of and adjustments to the relative value units under the physician fee schedule for calendar year 2002. Federal Register 2002;81(67):20681-20687.

8. Bowen B, Castillo M, Haughton VM. Practice guideline for the performance and interpretation of magnetic resonance spectroscopy of the brain. AHRQ. 2002:49-52.

9. Jordan HS, et al. Magnetic resonance spectroscopy for brain tumors. 2003 [cited. June 17] Available at: http://www.cms.hhs.gov/ncdr/trackingsheet.asp?id=52. No longer available online.]

10. Mark DH, et al. Magnetic resonance spectroscopy for evaluation of suspected brain tumor. TEC Assessment Program 2003 June 2003 [cited 2005 March 25]; Available from: http://www.bcbs.com/tec/vol18/18_01.html.

11. Phurrough SE, et al. Decision memo for magnetic resonance spectroscopy for brain tumors (CAG-00141N). Jan. 29, 2004 [cited 2005 March 25]; Available from: http://www.cms.hhs.gov/mcd/viewdecisionmemo.asp?id=52.

Dr. Ross is director of clinical MR, Mr. Lin is a clinical coordinator, and Ms. Hollis is a medical billing specialist, all at Huntington Medical Research Institutes in Pasadena, CA.