Dual-energy CT in the differentiation between adrenal adenomas and metastases: Usefulness of material density maps and monochromatic images.

OBJECTIVE: To evaluate the behavior of adrenal adenomas and metastases with dual-energy CT, analyzing the attenuation coefficient in monochromatic images at three different levels of energy (45, 70, and 140 keV) and the tissue concentrations of fat, water, and iodine in material density maps, with the aim of establishing optimal cutoffs for differentiating between these lesions and comparing our results against published evidence. MATERIALS AND METHODS: This retrospective case-control study included oncologic patients diagnosed with adrenal metastases in the 6-12 months prior to the study who were followed up in our hospital between January and June 2020. For each case (patient with metastases) included in the study, we selected a control (patient with an adrenal adenoma) with a nodule of similar size. All patients were studied with a rapid-kilovoltage-switching dual-energy CT scanner, using a biphasic acquisition protocol. We analyzed the concentration of iodine in paired water-iodine images, the concentration of fat in the paired water-fat images, and the concentration of water in the paired iodine-water and fat-water images, in both the arterial and portal phases. We also analyzed the attenuation coefficient in monochromatic images (at 55, 70, and 140 keV) in the arterial and portal phases. RESULTS: In the monochromatic images, in both the arterial and portal phases, the attenuation coefficient at all energy levels was significantly higher in the group of patients with metastases than in the group of patients with adenomas. This enabled us to calculate the optimal cutoffs for classifying lesions as adenomas or metastases, except for the arterial phase at 55 KeV, where the area under the receiver operating characteristic curve (AUC) for the estimated threshold (0.68) was not considered accurate enough to classify the lesions. For the arterial phase at 70 keV, the AUC was 0.76 (95% CI: 0.663‒0.899); the optimal cutoff (42.4 HU) yielded 92% sensitivity and 60% specificity. For the arterial phase at 140 keV, the AUC was 0.94 (95% CI: 0.894‒0.999); the optimal cutoff (18.9 HU) yielded 88% sensitivity and 94% specificity). For the portal phase at 55 keV, the AUC was 0.76 (95% CI: 0.663‒0.899); the optimal cutoff (95.4 HU) yielded 68% sensitivity and 84% specificity. For the portal phase at 70 keV, the AUC was 0.82 (95% CI: 0.757‒0.955); the optimal cutoff (58.4 HU) yielded 80% sensitivity and 84% specificity. For the portal phase at 140 keV, the AUC was 0.9 (95% CI: 0.834‒0.987); the optimal cutoff (16.35 HU) yielded 96% sensitivity and 84% specificity. In the material density maps, in the arterial phase, significant differences were found only for the iodine-water pair, where the concentration of water was higher in the group with metastases (1018.8 ±â€¯7.6 mg/cm3 vs. 998.6 ±â€¯8.0 mg/cm3 for the group with adenomas, p < 0.001). The AUC was 0.97 (95% CI: 0.893‒0.999); the optimal cutoff (1012.5 mg/cm3) yielded 88% sensitivity and 96% specificity. The iodine-water pair was also significantly higher in metastases (1019.7 ±â€¯12.1 mg/cm3 vs. 998.5 ±â€¯9.1 mg/cm3 in adenomas, p < 0.001). The AUC was 0.926 (95% CI: 0.807‒0.977); the optimal cutoff (1009.5 mg/cm3) yielded 92% sensitivity and 92% specificity. Although significant results were also observed for the fat-water pair in the portal phase, the AUC was insufficient to enable a sufficiently accurate cutoff for classifying the lesions. No significant differences were found in the fat-water maps or iodine-water maps in the arterial or portal phase or in the water-fat map in the arterial phase. CONCLUSIONS: Monochromatic images show differences between the behavior of adrenal adenomas and metastases in oncologic patients studied with intravenous-contrast-enhanced CT, where the group of metastases had higher attenuation than the group of adenomas in both the arterial and portal phases; this pattern is in line with the evidence published for adenomas. Nevertheless, to our knowledge, no other publications report cutoffs for this kind of differentiation in contrast-enhanced monochromatic images obtained in rapid-kilovoltage-switching dual-energy CT scanners, and this is the first new contribution of our study. Regarding the material density maps, our results suggest that the water-iodine pair is a good tool for differentiating between adrenal adenomas and metastases, in both the arterial and portal phases. We propose cutoffs for differentiating these lesions, although to our knowledge no cutoffs have been proposed for portal-phase contrast-enhanced images obtained with rapid-kilovoltage-switching dual-energy CT scanners.

Frequency of lipid-poor adrenal adenomas in magnetic resonance imaging examinations of the abdomen.

Objective: To estimate the frequency of lipid-poor adenomas (LPAs) in magnetic resonance imaging (MRI) examinations. Materials and Methods: We retrospectively investigated adrenal lesions on MRI examinations performed in a total of 2,014 patients between January 2016 and December 2017. After exclusions, the sample comprised 69 patients with 74 proven adenomas. Two readers (reader 1 and reader 2) evaluated lesion size, laterality, homogeneity, signal drop on out-of-phase (OP) images, and the signal intensity index (SII). An LPA was defined as a lesion with no signal drop on OP images and an SII < 16.5%. For 68 lesions, computed tomography (CT) scans (obtained within one year of the MRI) were also reviewed. Results: Of the 69 patients evaluated, 42 (60.8%) were women and 27 (39.2%) were men. The mean age was 59.2 ± 14.1 years. Among the 74 confirmed adrenal adenomas evaluated, the mean lesion size was 18.5 ± 7.7 mm (range, 7.0-56.0 mm) for reader 1 and 21.0 ± 8.3 mm (range, 7.0-55.0 mm) for reader 2 (p = 0.055). On the basis of the signal drop in OP MRI sequences, both readers identified five (6.8%) of the 74 lesions as being LPAs. When determined on the basis of the SII, that frequency was three (4.0%) for reader 1 and four (5.4%) for reader 2. On CT, 21 (30.8%) of the 68 lesions evaluated were classified as LPAs. Conclusion: The prevalence of LPA was significantly lower on MRI than on CT. That prevalence tends to be even lower when the definition of LPA relies on a quantitative analysis rather than on a qualitative (visual) analysis.

Objetivo: Estimar a frequência de adenomas pobres em lipídios (APLs) em exames de ressonância magnética (RM). Materiais e Métodos: Investigaram-se, retrospectivamente, as lesões adrenais em exames de RM realizados de janeiro de 2016 a dezembro de 2017. Um total de 2.014 pacientes foi submetido a exames abdominais e, após exclusões, 69 pacientes com 74 adenomas foram recuperados. Determinaram-se o tamanho da lesão, a lateralidade, a homogeneidade, a queda do sinal em imagens fora-de-fase (FF) e o índice de intensidade do sinal (IIS). Foram utilizadas as seguintes definições para APLs: sem queda de sinal nas imagens FF e IIS < 16,5%. Para 68 lesões, havia imagens de tomografia computadorizada (TC), com intervalo de até um ano da RM, que também foram analisadas. Resultados: Sessenta e nove pacientes foram incluídos, sendo 42 mulheres (60,8%) e 27 homens (39,2%). A média de idade foi 59,2 ± 14,1 anos. O tamanho médio do adenoma adrenal foi 18,5 ± 7,7 mm para o leitor 1 (7,0-56,0 mm) e 21,0 ± 8,3 mm (7,0-55,0 mm) para o leitor 2 (p = 0,055). A queda de sinal nas imagens FF mostrou que a frequência de APLs para ambos os leitores foi 6,8% (5/74). Para a análise quantitativa, a frequência foi 4,0% (3/74) para o leitor 1 e 5,4% (4/74) para o leitor 2. A frequência de APLs nas imagens de TC foi 21/68 lesões (30,8%). Conclusão: A prevalência de APLs em imagens de RM foi significativamente menor do que em exames de TC. Essa prevalência tende a ser ainda menor quando a definição de APL é baseada na análise quantitativa (IIS < 16,5%), em vez da análise visual.

Frequency of lipid-poor adrenal adenomas in magnetic resonance imaging examinations of the abdomen / Frequência de adenomas adrenais pobres em lipídios em exames de ressonância magnética abdominal

Abstract Objective: To estimate the frequency of lipid-poor adenomas (LPAs) in magnetic resonance imaging (MRI) examinations. Materials and Methods: We retrospectively investigated adrenal lesions on MRI examinations performed in a total of 2,014 patients between January 2016 and December 2017. After exclusions, the sample comprised 69 patients with 74 proven adenomas. Two readers (reader 1 and reader 2) evaluated lesion size, laterality, homogeneity, signal drop on out-of-phase (OP) images, and the signal intensity index (SII). An LPA was defined as a lesion with no signal drop on OP images and an SII < 16.5%. For 68 lesions, computed tomography (CT) scans (obtained within one year of the MRI) were also reviewed. Results: Of the 69 patients evaluated, 42 (60.8%) were women and 27 (39.2%) were men. The mean age was 59.2 ± 14.1 years. Among the 74 confirmed adrenal adenomas evaluated, the mean lesion size was 18.5 ± 7.7 mm (range, 7.0-56.0 mm) for reader 1 and 21.0 ± 8.3 mm (range, 7.0-55.0 mm) for reader 2 (p = 0.055). On the basis of the signal drop in OP MRI sequences, both readers identified five (6.8%) of the 74 lesions as being LPAs. When determined on the basis of the SII, that frequency was three (4.0%) for reader 1 and four (5.4%) for reader 2. On CT, 21 (30.8%) of the 68 lesions evaluated were classified as LPAs. Conclusion: The prevalence of LPA was significantly lower on MRI than on CT. That prevalence tends to be even lower when the definition of LPA relies on a quantitative analysis rather than on a qualitative (visual) analysis.

Resumo Objetivo: Estimar a frequência de adenomas pobres em lipídios (APLs) em exames de ressonância magnética (RM). Materiais e Métodos: Investigaram-se, retrospectivamente, as lesões adrenais em exames de RM realizados de janeiro de 2016 a dezembro de 2017. Um total de 2.014 pacientes foi submetido a exames abdominais e, após exclusões, 69 pacientes com 74 adenomas foram recuperados. Determinaram-se o tamanho da lesão, a lateralidade, a homogeneidade, a queda do sinal em imagens fora-de-fase (FF) e o índice de intensidade do sinal (IIS). Foram utilizadas as seguintes definições para APLs: sem queda de sinal nas imagens FF e IIS < 16,5%. Para 68 lesões, havia imagens de tomografia computadorizada (TC), com intervalo de até um ano da RM, que também foram analisadas. Resultados: Sessenta e nove pacientes foram incluídos, sendo 42 mulheres (60,8%) e 27 homens (39,2%). A média de idade foi 59,2 ± 14,1 anos. O tamanho médio do adenoma adrenal foi 18,5 ± 7,7 mm para o leitor 1 (7,0-56,0 mm) e 21,0 ± 8,3 mm (7,0-55,0 mm) para o leitor 2 (p = 0,055). A queda de sinal nas imagens FF mostrou que a frequência de APLs para ambos os leitores foi 6,8% (5/74). Para a análise quantitativa, a frequência foi 4,0% (3/74) para o leitor 1 e 5,4% (4/74) para o leitor 2. A frequência de APLs nas imagens de TC foi 21/68 lesões (30,8%). Conclusão: A prevalência de APLs em imagens de RM foi significativamente menor do que em exames de TC. Essa prevalência tende a ser ainda menor quando a definição de APL é baseada na análise quantitativa (IIS < 16,5%), em vez da análise visual.