Note from the American College of Radiology (ACR) and the National Guideline Clearinghouse (NGC): ACR has updated its Relative Radiation Level categories and Rating Scale. The Rating Scale now includes categories (1,2,3 = Usually not appropriate; 4,5,6 = May be appropriate; 7,8,9 = Usually appropriate). See the original guideline document for details.
ACR Appropriateness Criteria®
Clinical Condition: Headache
Variant 1: Chronic headache. No new features.
Radiologic Procedure |
Rating |
Comments |
RRL* |
MRI head without and with contrast |
4 |
See statement regarding contrast in text under "Anticipated Exceptions." |
None |
MRI head without contrast |
4 |
|
None |
CT head without contrast |
4 |
|
Med |
CT head without and with contrast |
4 |
|
Med |
MRA head with or without contrast |
2 |
|
None |
Arteriography cervicocerebral |
2 |
|
Med |
CTA head |
2 |
|
Med |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 2: Chronic headache with new features.
Radiologic Procedure |
Rating |
Comments |
RRL* |
MRI head without and with contrast |
8 |
See statement regarding contrast in text under "Anticipated Exceptions." |
None |
MRI head without contrast |
7 |
|
None |
CT head without contrast |
5 |
If new features highly suggestive of intracranial hemorrhage, see variant 3. If MRI unavailable or contraindicated. |
Med |
MRA head with or without contrast |
5 |
Selected cases when vascular disease suspected. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head without and with contrast |
4 |
If MRI unavailable or contraindicated. |
Med |
CTA head |
4 |
Not generally appropriate for screening or first study. To be used in combination with CT. For problem solving. |
Med |
Arteriography cervicocerebral |
2 |
Not used as a primary diagnostic tool. |
Med |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 3: Sudden onset of severe headache ("Worst headache of one's life," "thunderclap headache").
Radiologic Procedure |
Rating |
Comments |
RRL* |
CT head without contrast |
9 |
|
Med |
CTA head |
8 |
Usage of CT versus MRI depends on local preference and availability. |
Med |
MRA head with or without contrast |
8 |
Usage of CT versus MRI depends on local preference and availability. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
Arteriography cervicocerebral |
7 |
|
Med |
MRI head without contrast |
7 |
May be helpful after CT depending on CT findings. |
None |
MRI head without and with contrast |
6 |
May be helpful after CT depending on CT findings. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head without and with contrast |
6 |
|
Med |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 4: Sudden onset of unilateral headache, or suspected carotid or vertebral dissection or ipsilateral Horner's syndrome.
Radiologic Procedure |
Rating |
Comments |
RRL* |
CTA head and neck |
8 |
Usage of CT versus MRI depends on local preference and availability. |
Med |
MRA head and neck with or without contrast |
8 |
Usage of CT versus MRI depends on local preference and availability. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
MRI head without and with contrast |
8 |
With diffusion-weighted sequences. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
MRI head without contrast |
8 |
With diffusion-weighted sequences. |
None |
CT head without contrast |
7 |
|
Med |
Arteriography cervicocerebral |
7 |
|
Med |
CT head without and with contrast |
6 |
|
Med |
US carotid duplex |
3 |
|
None |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 5: Headache, suspected intracranial complication of sinusitis and/or mastoiditis. (See the National Guideline Clearinghouse [NGC] summary of the ACR Appropriateness Criteria® sinonasal disease.)
Radiologic Procedure |
Rating |
Comments |
RRL* |
MRI head without and with contrast |
8 |
See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head without contrast |
7 |
Sinus imaging may also be indicated. |
Med |
MRI head without contrast |
7 |
|
None |
CT head without and with contrast |
6 |
Sinus imaging may also be indicated. |
Med |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 6: New headache in patient older than 60. Sedimentation rate higher than 55, temporal tenderness. Suspected temporal arteritis.
Radiologic Procedure |
Rating |
Comments |
RRL* |
MRI head without contrast |
8 |
With diffusion-weighted sequences. |
None |
MRI head without and with contrast |
7 |
With diffusion-weighted sequences. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head without contrast |
6 |
|
Med |
CTA head and neck |
5 |
|
Med |
MRA head and neck with or without contrast |
5 |
See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head without and with contrast |
5 |
|
Med |
Arteriography cervicocerebral |
4 |
If noninvasive imaging unrewarding. |
Med |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 7: New headache in human immunodeficiency virus (HIV)-positive or immunocompromised individual.
Radiologic Procedure |
Rating |
Comments |
RRL* |
MRI head without contrast |
8 |
|
None |
MRI head without and with contrast |
8 |
See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head without and with contrast |
6 |
|
Med |
CT head without contrast |
5 |
If MRI not available. |
Med |
MRA head with or without contrast |
5 |
Indicated if vascular lesion suspected. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CTA head |
5 |
Indicated if vascular lesion suspected. |
Med |
FDG-PET head |
4 |
Useful if indeterminate mass present. |
High |
Tc-99m HMPAO SPECT head |
2 |
Useful if indeterminate mass present. |
High |
Arteriography cervicocerebral |
2 |
If noninvasive imaging unrewarding. |
Med |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 8: New headache in pregnant patient.
Radiologic Procedure |
Rating |
Comments |
RRL* |
MRI head without contrast |
8 |
Usage of CT versus MRI depends on local preference and availability. |
None |
CT head without contrast |
8 |
Usage of CT versus MRI depends on local preference and availability. |
Med |
MRI head without and with contrast |
5 |
Pregnancy is a relative contraindication to gadolinium administration. Reserve for urgent medical necessity. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
MRA head with or without contrast |
5 |
MR venography (MRV) should also be performed. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head with contrast |
3 |
For urgent medical necessity only. |
Med |
CTA head |
2 |
If MRI not available, contraindicated, or inconclusive. CTV may also be performed. |
Med |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 9: New headache. Suspected meningitis/encephalitis.
Radiologic Procedure |
Rating |
Comments |
RRL* |
MRI head without and with contrast |
8 |
Usage of CT versus MRI depends on local preference and availability. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head without contrast |
8 |
Usage of CT versus MRI depends on local preference and availability. To exclude signs of increased intracranial pressure, mass, or mass effect. |
Med |
MRI head without contrast |
6 |
|
None |
MRA head with or without contrast |
6 |
MRV should also be performed. See statement regarding contrast in text under "Anticipated Exceptions." |
None |
CT head without and with contrast |
6 |
MRI preferable, depending on availability. |
Med |
CTA head |
3 |
Useful for problem solving or if there is a strong suspicion of vascular disease. |
Med |
Rating Scale: 1=Least appropriate, 9=Most appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Summary of Literature Review
Headache is one of the most frequent ailments of the human race. Studies of the prevalence of headache of any kind in populations have estimated frequencies of 11% to 48% in children and 6% to 71% in adults. As with migraine, age, gender, and case definition may largely account for this variance. However, a higher prevalence of headache has been found by surveys in Europe and North America than by those of Asian and South American countries. A survey of the Canadian population showed that only about 20% of people there are headache free. Prevalence of migraine shows a clear-cut gender difference, as well as genetic factors, affecting about 15% to 18% of women and 6% of men. It occurs most commonly in men and women 25 to 55 years of age. Muscle contraction or tension accounts for most of the nonmigraine headaches encountered in population surveys.
By comparison, the frequency of pathology that can present with headache is rather low. The yearly incidence of brain tumors in the United States is 46 per 100,000. For subarachnoid hemorrhage, the yearly incidence is nine per 100,000. Arteriovenous malformations (AVMs) are about one-tenth as frequent as saccular aneurysms. Only a subset of these patients presents with isolated headache. In a retrospective review of the presentation of 111 brain tumors, headaches were a symptom in 48%, equally for primary and metastatic brain tumors. Headaches were similar to tension type in 77%, migraine type in 9%, and other types in 14%. The typical headache was bifrontal but worse ipsilaterally and was the worst symptom in only 45% of patients. Other studies have found a higher frequency, but sometimes the headache preceded the diagnosis of brain tumor by several years, bringing up the possibility of an association with this common complaint, rather than causality. In children with brain tumors, headache was present in 62%, more often with infratentorial tumors. Because tumors are rare and only about half of them present with headache, it becomes apparent that if all patients with headache undergo imaging procedures, a large proportion of the studies will be negative.
Several studies have confirmed the low yield of imaging procedures in individuals presenting with isolated headache—that is, headache unaccompanied by other neurological findings. Most of them are retrospective reviews. The patients were referred for imaging because the referring physician suspected pathology detectable by imaging or the patients requested the study to be certain that they did not have a brain tumor. A prospective review of 293 computed tomography (CT) scans ordered in an ambulatory family practice setting disclosed that most of them were ordered because the clinician suspected that a tumor (49%) or a subarachnoid hemorrhage (SAH) (9%) might be present. Fifty-nine (17%) were ordered because of patient expectation or medico-legal concerns.
Computed Tomography and Magnetic Resonance Imaging
Studies before 1991 on the yield of CT or magnetic resonance imaging (MRI) in patients with headache but normal neurological examination were reviewed by Frishberg. The scans examined in most of the larger studies were performed with first-generation CT equipment. In addition, Frishberg included three more studies in his excellent meta-analysis. Of 897 studies in patients with migraine, only four were positive, three for a tumor and one for an AVM, giving a 0.4% yield of potentially treatable lesions. In patients with unspecified headache, 1,825 scans yielded a total of 43 lesions (21 tumors, 8 hydrocephalus, 6 AVMs, 5 subdural hematomas, and 3 aneurysms), for a 2.4% yield of potentially treatable lesions. However, two studies in this group were performed at tertiary referral centers (the Mayo Clinic and the Cleveland Clinic) in the early days of CT and had a 500% higher rate of clinically important findings than more recent prospective studies. If these two studies are not included among those performed in patients with unspecified headache, the total number of potentially treatable lesions is reduced to three in 725 studies (0.4%). A potential bias for the early series, however, is that the studies were performed with first-generation equipment, which was likely to have less sensitivity than currently used units.
Of 1,999 scans reported in other series, mostly using CT, only 21 (1%) disclosed treatable lesions. Most of the positive cases occurred in the series of Becker et al, which included an unspecified number of patients with abnormal neurological findings. If this series is excluded from the analysis, only nine out of 1,999 patients (0.5%) had treatable findings. In a retrospective review of charts from 1,074 consecutive emergency department patients who underwent cranial CT, headache was associated with low yield of abnormality.
When considering such a common disorder as headache, the indications for the use of imaging procedures become particularly relevant. This is particularly true in the face of emerging and rapidly evolving technologies in use today. In frequent conditions, performing low-yield studies is more likely to result in false positive results, with the consequent risk of causing additional and unnecessary procedures to be performed. As indicated above, the yield of positive studies in patients referred with isolated, nontraumatic headache is about 0.4%. In terms of cost, if one assumes the cost of a CT scan to be $400 and of an MRI to be $900, to detect a lesion with CT would cost $100,000 and with MRI, $225,000.
One should not assume, however, that there is no social benefit in negative imaging studies in the setting of headache. Indeed, headache symptoms can be quite ominous and onerous to the one suffering them, and there can be tremendous costs with respect to productivity and quality-of-life issues. Moreover, health care providers perceive value in imaging headache when the fear of litigation is accounted for. While it is beyond the scope of this review to assess the factors and inherent value of negative imaging tests in headache imaging, it must be emphasized that costs of detection or screening in imaging headache are always overstated when the value of negative results is not factored into the analysis.
Thunderclap Headache
Some headache presentations require further discussion. A patient presenting with a sudden, severe headache ("the worst headache of my life," "thunderclap headache"), particularly if it is not a migraine or if the pattern of the headache is clearly different from the patient's usual headaches, is at a significantly higher risk of having an SAH, which is more often related to an aneurysm than to an AVM. In a combination of three series, as many as 165 of 350 patients (47%) presenting with thunderclap headache had an SAH. If the CT scan is negative, a lumbar puncture should be performed to disclose additional instances of SAH. These patients may require magnetic resonance angiography (MRA), computed tomography angiography (CTA), and/or catheter angiography to determine the nature and location of the lesion.
Severe Unilateral Headache
Sudden, severe unilateral headache in a young patient, particularly when it radiates into the neck and is accompanied by ipsilateral Horner's syndrome, may be the result of arterial dissection of the carotid or vertebral arteries. In a series of 161 patients, headache was reported by 68% of them, and, when present, it was the initial manifestation in 47% of those with carotid dissection and in 33% of those with vertebral dissection. Although some of these patients had stroke-like syndromes, others did not, or they developed them several days after an initial presentation with isolated headache. The pattern of the pain's radiation will often differ enough to make the patient suspect that this is not a regular headache. In this case, MRI, MRA, CTA, and/or catheter angiography are particularly useful to identify the nature of the lesion. Current practice is to anticoagulate these patients to prevent thrombosis at the site of the stenotic lesion. For this reason, identification of the pathology is important.
In 315 children with isolated headache scanned at Boston Children's Hospital, 4% had surgical space-occupying lesions. Sleep-related headache and no family history of migraine were the strongest predictors. (See the NGC summary of the ACR Appropriateness Criteria® headache - child.) The comments made above about selected populations referred to tertiary care centers apply to this example also.
Temporal Headache
Patients older than 55 years with new-onset headache in the temple regions, particularly when they have tender superficial temporal arteries, should be studied for temporal arteritis. Treatment with steroids may forestall vision loss or brainstem strokes.
Pregnant Patients
Patients presenting with headache in pregnancy may also have higher yields of a pathologic etiology. While guidance in this area is still limited, these patients appear to require careful consideration regarding neuroimaging, as a recent study found a 27% underlying etiology for headache in pregnant patients presenting to the emergency department.
Intracranial Pathology
New-onset headache in populations predisposed to intracranial pathology also results in a much higher yield of findings by CT or MRI. For instance, a series of 49 HIV-positive individuals had an 82% yield of positive pathology. Although cryptococcal meningitis was most common (39%), toxoplasmosis was a close second (16%), and a number of patients had other mass lesions identified by CT. Patients with known cancer should also be scanned when a headache develops or changes in characteristics. In the Andes population, the rate of headache is low, whereas cysticercosis is common. As a result, CT of patients with headache yielded a 33% rate of positive studies. Although transcranial ultrasound (US), single photon emission CT (SPECT), and positron emission tomography (PET) are not currently routinely used in the evaluation of patients with headache, PET and SPECT may be useful in differentiating tumor from infection in high-risk patients.
Summary
- Screening patients with isolated, nontraumatic headache by means of CT or MRI is usually not warranted, but in some cases imaging may be required depending on the clinical circumstances.
- For some types of headache or populations at risk, these procedures are more likely to be positive.
- Thunderclap headaches, headaches radiating to the neck, and temporal headaches in an older individual are examples of headaches for which imaging procedures may be helpful.
- Patients with suspected meningitis and those presenting with headaches in pregnancy also often pose important diagnostic challenges.
- Human immunodeficiency virus (HIV)-positive individuals, cancer patients, or other populations at high risk of intracranial disease also should be screened when presenting with new-onset headaches.
Anticipated Exceptions
Nephrogenic systemic fibrosis (NSF) is a disorder with a scleroderma-like presentation and a spectrum of manifestations that can range from limited clinical sequelae to fatality. It appears to be related to both underlying severe renal dysfunction and the administration of gadolinium-based contrast agents. It has occurred primarily in patients on dialysis, rarely in patients with very limited glomerular filtration rate (GFR) (i.e., <30 mL/min/1.73 m2), and almost never in other patients. There is growing literature regarding NSF. Although some controversy and lack of clarity remain, there is a consensus that it is advisable to avoid all gadolinium-based contrast agents in dialysis-dependent patients unless the possible benefits clearly outweigh the risk, and to limit the type and amount in patients with estimated GFR rates <30 mL/min/1.73 m2. For more information, please see the ACR Manual on Contrast Media (see the "Availability of Companion Documents" field).
Abbreviations
- CT, computed tomography
- CTA, computed tomography angiography
- FDG-PET, fluorine-18-2-fluoro-2-deoxy-D-glucose positron emission tomography
- HMPAO, hexamethylpropyleneamine oxime
- Med, medium
- MRA, magnetic resonance angiography
- MRI, magnetic resonance imaging
- SPECT, single photon emission computer tomography
- US, ultrasound
Relative Radiation Level |
Effective Dose Estimate Range |
None |
0 |
Minimal |
<0.1 mSv |
Low |
0.1-1 mSv |
Medium |
1-10 mSv |
High |
10-100 mSv |