Brain PET/MRI: Differentiating between recurrent metastasis and radiation necrosis
This case focuses on a 51-year-old woman initially diagnosed with endometrial carcinoma in October 2015. Surgical pathology appeared consistent with endometrioid adenocarcinoma (architectural grade 1, nuclear grade 2 to 3, Federation of Gynecology and Obstetrics grade 2).
The patient developed metastases to both breasts and ovaries in 2016. She underwent total abdominal hysterectomy, bilateral salpingo-oophorectomy and chemotherapy (paclitaxel, followed by paclitaxel protein-bound particles for injectable suspension [Abraxane, Celgene] and carboplatin).
The patient subsequently received immunotherapy — including infliximab (Remicade, Janssen) from May 2017 to June 2017, and vedolizumab (Entyvio, Takeda) from August 2017 through October 2017 — as well as prednisone. Treatment was complicated by steroid-induced myopathies.
In November 2017, the patient developed neurologic deficits and was diagnosed with a single right frontal brain metastasis. She underwent stereotactic radiosurgery on Nov. 15, 2017.
She presented to the ED in late January of this year with a progressive decline in speech over 1 to 2 weeks, as well as acute right hemiparesis.
Brain MRI obtained upon presentation showed significant increase in size of a necrotic lesion in the right precentral gyrus suggestive of tumor progression. The patient began corticosteroid therapy, and she and her care team decided to pursue hospice care only.
However, while on high-dose steroids, the patient significantly improved back to her recent baseline. Given her significant improvement on steroids, the short time since stereotactic radiosurgery and imaging findings, it became apparent the lesion may have represented radiation necrosis.
Of note, the patient had no evidence of disease systemically and her cancer was otherwise controlled. Therefore, she underwent 18F-fluorodeoxyglucose (FDG) PET/MRI with perfusion-weighted MRI for further evaluation.
Brain MRI prior to stereotactic radiosurgery obtained on Nov. 15, 2017, demonstrated a well-circumscribed intra-axial enhancing mass in the left frontoparietal perirolandic region. The mass, measured 2.9 cm by 2.7 cm, was concerning for intracranial metastasis (Figure 1).
Perilesional edema with mass effect was visible on adjacent sulci, but there was no midline shift or transtentorial herniation.
Following radiosurgery, brain MRI obtained on Dec. 12, 2017, demonstrated expected interval contraction of the previously described peripherally enhancing metastatic lesion in the left postcentral gyrus measuring 1.1 cm by 1.7 cm by 1.4 cm (anteroposterior by transverse x craniocaudal). By this point, the associated surrounding edema had nearly completely resolved (Figure 2).
Subsequent brain MRI upon presentation to the ED on Jan. 31 demonstrated significant interval enlargement of the peripherally enhancing subcortical/cortical mass centered in the lateral left postcentral gyrus, now measuring 1.7 cm by 2.8 cm by 2.3 cm.
Imaging showed extensive surrounding vasogenic edema involving the left centrum semiovale, as well as left frontal and parietal lobes, extending into the posterior left internal and external capsules. There also was associated mass effect, including effacement of the third ventricle and a 3-mm rightward midline shift as measured at the level of the foramen of Monro (Figures 3a and 3b).
Given the patient’s marked clinical improvement over the subsequent weeks, a dedicated brain FDG PET/MRI hybrid imaging study was performed on April 12 to evaluate for metabolic, dynamic perfusion and permeability, as well as structural evolution of the lesion, to better ascertain whether the findings truly represented progression of intracranial metastatic disease.
The patient underwent PET imaging from the vertex of the skull to the foramen magnum on a 3 T mMR scanner. Simultaneous MRI sequences with different contrasts and imaging planes were obtained of the head before and after IV administration of 4.4 mL gadobutrol (Gadavist, Bayer HealthCare). Structural MRI again demonstrated the centrally necrotic lesion with a thick rind of peripheral enhancement in the left postcentral gyrus, similar in size with prior imaging but suggestive of mildly increasing associated contrast enhancement. (Figures 4a and 4b).
Dynamic susceptibility contrast T2* perfusion analysis demonstrated hyperfusion with relative cerebral blood volume measure up to 4.7 mL per 100 g in areas of abnormal enhancement compared with contralateral normal-appearing white matter (relative cerebral blood volume of 1.2 mL per 100 g), with associated rapid wash-in and progressive wash-out within the lesion (Figure 5). Similarly, dynamic contrast-enhanced T1-weighted magnetic resonance analysis revealed focal markedly increased vascular leakiness — or permeability — suggestive of recurrent or progressive tumor (Figures 6a and 6b).
Metabolic images further corroborated these findings, with fused PET/MRI images demonstrating focal, markedly increased FDG uptake associated with the peripherally enhancing lesion in the left postcentral gyrus (maximum standard uptake value, 13.1), compatible with progression of brain parenchymal metastasis (Figures 7a, 7b and 7c). No additional hypermetabolic lesions were observed elsewhere.
Normal distribution of FDG activity was seen throughout the remaining cerebral hemispheres, basal ganglia, thalami, brainstem and cerebellum (Figure 8).
Radiation necrosis is a well-known phenomenon in the setting of brain radiation treatment.
Radiation-induced changes in the brain parenchyma can be categorized based on time of presentation as acute, early-delayed and late-delayed reactions.
Acute radiation effects typically include transient vasogenic edema within 12 to 48 hours following treatment, which is fully reversible without permanent neurologic sequelae.
Early-delayed postradiation changes develop within a few weeks to several months. They are related to temporary demyelination and vascular damage, which may be either fully or partially reversible, or may progress to permanent damage. Late-delayed reactions typically develop from several months to many years after treatment; they are irreversible and frequently are progressive.
In the subacute phase, it often is difficult to distinguish radiation therapy changes from recurrence based on clinical and conventional imaging characteristics alone. Typically, extensive tumor and peritumoral edema develops related to radiation effects. Similarly, contrast enhancement at this time — particularly in the tumor perimeter — reflects a host reactive response and not tumor recurrence.
Lesions that undergo stereotactic radiosurgery constitute a diagnostic challenge. In conventional follow-up MRI, a new ring-shaped contrast enhancement can arise at the site of the highest delivered dose as indication of blood-brain barrier disruption. Therefore, interpretation of findings on structural MRI — changing patterns of contrast enhancement and alteration of signal intensity on T2-/FLAIR-weighted images — is markedly limited, making differentiation of local tumor recurrence from radiation-induced changes difficult.
Dynamic contrast-enhanced imaging techniques using MRI or CT have been used to obtain measures of tumor vascular physiology and hemodynamics.
Many of the perfusion parameters have been correlated with tumor grade, aggressiveness and prognosis. Hypoxia or hypoglycemia — occurring in rapidly growing tumors — increases the expression of VEGF, which is not only a potent angiogenic factor but also a potent permeability factor.
Therefore, dynamic susceptibility contrast T2* magnetic resonance perfusion of highly proliferative neoplasms characteristically demonstrates increased relative cerebral blood volume with associated rapid wash-in and progressive washout kinetics. Conversely, with radiation injury, dynamic susceptibility contrast T2* perfusion maps demonstrate low mean transit time and low relative cerebral blood volume.
Tumor vessel permeability estimates based on dynamic contrast-enhanced T1-weighted magnetic resonance perfusion techniques — which are not affected by susceptibility artifacts — can help obtain a better estimation of vascular leakiness.
Therefore, higher vascular leakiness on dynamic contrast enhanced T1-weighted magnetic resonance perfusion Ktrans maps suggests a recurrent or progressive tumor, whereas areas of radiation necrosis typically demonstrate low permeability.
Metabolic imaging, specifically PET with 18F-FDG, is a well-established modality widely used for staging and restaging of malignancy. Studies have suggested that molecular imaging techniques may help overcome limitations in assessment of intracranial metastases that have previously undergone focused high-dose radiotherapy, especially in clinically equivocal situations.
18F-FDG PET imaging is based on increased glucose utilization by malignant cells, with degree of uptake (ie, standard uptake values) correlating closely to histologic grade. In contrast, if evaluating tumor recurrence vs. radiation necrosis, 18F-FDG PET demonstrates hypometabolism in the region of radiation injury. However, low-grade gliomas may occasionally be hypometabolic at baseline, therefore leading to false-negative results.
Novel amino acid-based PET radiotracers have been particularly useful in neuro-oncology because, unlike FDG, the background uptake of amino acids in normal brain parenchyma is relatively low, providing good contrast with tumor uptake.
The mechanism of accumulation of L-methyl-11C-methionine (11C-MET) by cells remains unclear. Postulated hypotheses include increased protein synthesis by proliferative cells, active carrier-mediated transport across the tumor cell membrane, disruption of the blood-brain barrier and high vascular density in neoplastic tissue.
11C-MET and O-(2-18F-Fluoroethyl)-L-Tyrosine have reported sensitivity values in the range of 78% and specificity values in the range of 100% among patients after stereotactic radiosurgery.
Therefore, amino acid PET allows for sensitive monitoring of treatment response, early detection of tumor recurrence and improved differentiation of tumor recurrence (ie, high MET uptake) from radiation change (ie, uptake similar to contralateral cortex).
In conclusion, differentiation between radiation-induced lesions and tumor recurrence after focused high-dose radiotherapy of brain metastases is challenging.
PET imaging provides additional information about tumor metabolism. This allows a more accurate assessment, especially in clinically equivocal situations, particularly when radiolabeled amino acids are used.
Systematic addition of high-resolution MRI information to PET data provides accurate and consistent information about underlying structures, helps overcome difficulties in anatomic localization on PET images, may exclude or identify the presence of multiple pathologies, and improves scan interpretation without additional radiation safety concerns.
Galldiks N and Langen KJ. Front Neurol. 2016;doi:10.3389/fneur.2016.00120.
Galldiks N, et al. Neuroimage Clin. 2016;doi:10.1016/j.nicl.2016.12.020.
Glaudemans AW, et al. Eur J Nucl Med Mol Imaging. 2013;10.1007/s00259-012-2295-5.
Jain R, et al. J Neurooncol. 2010;doi:10.1007/s11060-010-0139-3.
Kickingereder P, et al. Radiat Oncol. 2013;doi:10.1186/1748-717X-8-52.
Kits A, et al. Ann Nucl Med. 2018;doi:10.1007/s12149-017-1227-7.
Minamimoto R, et al. PLoS One. 2015;doi:10.1371/journal.pone.0132515.
Romagna A, et al. Radiat Oncol. 2016;11:139.
Shin HK, et al. J Korean Neurosurg Soc. 2016;doi:10.3340/jkns.2016.59.4.392.
Tasuyuguchi N, et al. J Neurosurg. 2003;98:1056-1064.