Introduction

Among primary brain tumors, craniopharyngiomas (2.5 %) and germ cell tumors (GCTs) (2.3 %) are common in the Japanese population [1]. The epidemiology, markers, and locations of the tumors are useful for diagnosing both craniopharyngiomas and GCTs. These lesions, however, are likely to occur in the suprasellar region in young patients and exhibit various findings on magnetic resonance (MR) imaging. The titer of human chorionic gonadotropin-beta (HCG-β) in the cerebral spinal fluid, which generally indicates the presence of GCTs, is often elevated in patients with craniopharyngiomas [2, 3]. On the other hand, there are many GCTs without elevation of the titer of HCG-β and alpha-fetoprotein [4, 5]. In such cases, accurately differentiating between craniopharyngiomas and GCTs preoperatively is very difficult. However, the treatment strategies for these tumors differ (craniopharyngiomas should be removed surgically and GCTs should be treated with chemotherapy); therefore, it is essential to accurately distinguish between craniopharyngiomas and GCTs preoperatively.

Diffusion-weighted imaging (DWI) and the apparent diffusion coefficient (ADC) are used to diagnose intracranial tumors and predict the histopathological grade of the tumor [6–9], and the advantages of high b value DWI have been reported [8, 10–13]. In the present study, we retrospectively evaluated the efficacy of DWI and ADC values at b = 1000 and b = 4000 s/mm2 on 3T MR imaging and assessed the usefulness of DWI and the ADC values for differentiating between craniopharyngiomas and GCTs.

Materials and methods

Patients

A total of 55 patients who underwent 3T MR imaging and DWI scans to detect tumors at Hiroshima University Hospital between September 2006 and December 2013 were included in this retrospective study. All 55 patients were treated surgically at our institution and received a histological diagnosis; however, we excluded 12 patients with cystic lesions from this study in order to prevent the partial volume effect on the measurement of regions of interest (ROIs). The median age of the 19 patients with craniopharyngioma and 24 patients with GCT at the time of surgery was 20 years (range 5 to 70; Table 1). The GCTs included several histopathological types: 16 patients with pure germinoma, 1 patient with mature teratoma, 1 patient with immature teratoma, 1 patient with choriocarcinoma, 2 patients with yolk sac tumors + immature teratomas, 1 patient with germinoma + immature teratoma, 1 patient with germinoma + mature teratoma, and 1 patient with germinoma + choriocarcinoma. The patients were categorized into two groups: those with pure germinoma (n = 16) and those with non-pure germinoma (n = 8). Calcification on computed tomography (CT) was founded in the craniopharyngiomas (n = 11) and GCTs (n = 17). The focal calcification of GCTs included the pineal grand calcification in a maximum of 13 patients.

Table 1 Baseline characteristics of the patients
Full size table

MR imaging

MR images were acquired on a 3T superconducting system (Signa Horizon; Medical Systems, Milwaukee, WI, USA). Axial DWI was performed with b values of 1000 and 4000 s/mm2, an effective gradient registered at 40 mT/m and a slew rate of 150 mT/m/ms. The parameters at b = 1000 and b = 4000 s/mm2 using an eight-channel-phased array head coil at repetition time (TR) 5000 ms with echo times (TEs) of 73.7 and 101.1 ms were as follows: field of view (FOV) = 22 × 22 cm, section thickness = 6 mm, intersection gap = 1.0 mm, number of slices = 20, and data acquisition matrix = 128 × 128. Tumors located in the intrasellar or suprasellar region were evaluated using periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) DWI to reduce susceptibility artifacts [7, 14]. For axial PROPELLER DWI, the b value was 1000 s/mm2 with the following settings: TR = 5000 ms, TE = 125 ms, FOV = 22 × 22 cm, section thickness = 4 mm, intersection gap = 0.3 mm, and data acquisition matrix = 128 × 128. The patients were subsequently divided into the following three categories for the evaluation of the signal intensity on DWI: (1) low intensity, (2) iso intensity, and (3) high intensity.

All ADC maps generated with the software program (Functool; GE Medical Systems) were obtained by calculating the signal intensity on DWI at two different b values (0 and 1000 s/mm2 for ADC map b = 1000, and 0 and 4000 s/mm2 for ADC map b = 4000 s/mm2 on a pixel-by-pixel basis). The ADC was determined by manually placing the ROIs in the respective tumor regions on ADC maps of b = 1000 s/mm2, and the procedure was repeated with an approximately similar size of ROI placed at the same site on the ADC maps of b = 4000 s/mm2. ROIs were purposely placed over the gadolinium (Gd)-enhanced components of the tumors in order to avoid volume averaging with cystic, hemorrhagic, or calcified regions that may influence the ADC values. The number of ROIs appropriated depended on the size of the enhanced lesion. Based on 1–3 ROIs on the ADC maps at b = 1000 and b = 4000 s/mm2, the ADC value for each tumor was calculated and expressed as the minimum (ADCMIN), mean (ADCMEAN), and maximum (ADCMAX) absolute values.

Statistical analysis

All statistical analyses were performed using the software package SPSS 16.0J for Windows (SPSS Inc., Chicago, IL, USA). The relationships between the average ADC values were evaluated using a logistic discriminant analysis with the model, Pr (Y = 1|x) = 1 / {1 + exp (−β0 − β1 x)} where Y denotes the binary variable having a value of 1 in cases of GCTs and 0 in case of craniopharyngiomas and x denotes the variable of the ADC measurement. The chi-square test and Fisher’s exact test were used for comparisons among the categories of signal intensity on DWI. The values are expressed as medians. Differences are represented as the t value of the magnitude of estimated coefficients β1 and were considered significant at a p value of <0.05.

Results

DWI

The craniopharyngiomas (Fig. 1) showed low intensity on DWI at b = 1000 s/mm2 (n = 14, 73.6 %), while the GCTs (Fig. 2) showed high intensity (n = 17, 70.8 %) (p = 0.0058). The differences were not influenced by the histopathological diagnosis of the adamantinomatous or papillary type or the presence/absence of pure germinoma. DWI at b = 4000 s/mm2 demonstrated similar findings; low intensity for craniopharyngiomas (n = 10, 83.3 %) and high intensity for GCTs (n = 18, 78.3 %) (p = 0.0009) (Fig. 3). The differences in intensity between the craniopharyngiomas and GCTs were clear on DWI at b = 4000 s/mm2.

Fig. 1
figure 1

Magnetic resonance images of a craniopharyngioma in a 56-year-old female (a–c). A suprasellar mass was detected on a gadolinium-enhanced T1-weighted image (a), which showed low intensity on DWI at b = 1000 s/mm2 (b) and at b = 4000 s/mm2 (c)

Full size image
Fig. 2
figure 2

Magnetic resonance images of a pure germinoma in a 29-year-old male (a–c). A suprasellar mass was observed on a gadolinium-enhanced T1-weighted image (a), which showed slightly high intensity on DWI at b = 1000 s/mm2 (b) and high intensity at b = 4000 s/mm2 (c). The high intensity of the pure germinoma was emphasized on DWI at b = 4000 s/mm2

Full size image
Fig. 3
figure 3

In this study, the craniopharyngiomas were likely to exhibit low intensity, while the germ cell tumors displayed high intensity, on DWI at b = 1000 and b = 4000 s/mm2. The differences in intensity between the craniopharyngiomas and GCTs were more remarkable at b = 4000 than at b = 1000 s/mm2

Full size image

ADC values in the craniopharyngiomas and GCTs

The ADC values of the adamantinomatous- and papillary-type lesions are shown in Table 2. There were no significant differences in the ADC values between these tumors, although the papillary-type lesions demonstrated lower ADC values than the adamantinomatous-type lesions on all images. In addition, no significant differences were found in the ADC values between the pure germinoma and non-pure germinoma groups.

Table 2 ADC values in the subtypes of craniopharyngiomas and germ cell tumors
Full size table

ADC values in the calcified and non-calcified tumors

The ADC values of the calcified and non-calcified craniopharyngiomas were shown in Table 3. There was no significant difference in the ADC values between the two types of craniopharyngiomas, although the calcified craniopharyngiomas demonstrated lower ADC values than the non-calcified craniopharyngiomas. In the ADC values between the calcified and non-calcified GCTs, no significant differences were also found.

Table 3 ADC values in the craniopharyngiomas and germ cell tumors with or without calcification
Full size table

Differences between the ADC values of the craniopharyngiomas and GCTs

The craniopharyngiomas demonstrated clearly high ADC values in comparison with the GCTs on all images (Table 4). The logistic discriminant analysis of 34 patients with ADC values at b = 1000 and b = 4000 s/mm2 showed the advantage of the ADC values at b = 4000 s/mm2, especially ADCMIN at b = 4000 s/mm2 (p = 0.00172, t = −3.14). Using ADCMIN at b = 4000 s/mm2, the sensitivity and specificity for the differential diagnosis of GCT were 90.9 and 91.7 %, respectively.

Table 4 ADC values in the craniopharyngiomas and germ cell tumors
Full size table

Discussion

DWI and the ADC are often used to diagnose intracranial tumors [7, 9, 11, 15, 16]. It has also been reported that the ADC values of craniopharyngiomas are significantly higher than the ADC values of pituitary adenomas and meningiomas [9], while the ADCMIN values of craniopharyngiomas are lower than those of Rathke’s cleft cysts [7]. The differences between craniopharyngiomas and GCTs have been examined in a small number of cases, with no significant differences having been founded [9]. Therefore, this is the first study to demonstrate the differences in DWI and ADC values between craniopharyngiomas and GCTs.

High-grade tumors in pediatric patients as well as high-grade gliomas and high-grade meningiomas have been reported to exhibit lower ADC values than low-grade tumors [8, 9, 17]. It has also been reported that the mechanisms underlying the restricted diffusion of high-grade tumors originate in the high cellularity of these lesions [11, 18, 19]. From the viewpoint of microscopic findings, it is reasonable that the ADC values of malignant GCTs are lower than those of benign craniopharyngiomas, as GCTs usually display a high cellular pattern, while craniopharyngiomas usually show a loose, low cellular pattern. Craniopharyngiomas are rich in the non-cellular components of cholesterin, keratin, and microcysts (adamantinomatous type) in addition to villous fibrovascular stroma (papillary type), all of which may demonstrate relatively high ADC values. On the other hand, germinomas, the common component of GCTs in this study, are composed of uniform cells with large nuclei and poor stroma exhibiting lymphocytic infiltration that may demonstrate low ADC values.

In the present study, the advantage of ADCMIN at b = 4000 s/mm2 for differentiating between craniopharyngiomas and GCTs was demonstrated. Furthermore, it has been reported that high b value DWI is useful for predicting the histopathological grade of cerebral gliomas [10, 12]. Moreover, the ADC at b = 4000 s/mm2 displays a better inverse correlation reflective of the degree of tumor cellularity than the ADC value at b = 1000 s/mm2 [11]. Our findings coincide well with these findings of past reports.

Craniopharyngiomas have pathologically two subtypes, while GCTs have five. In addition, the mixed type of GCT exhibits structural heterogeneity in the tumor. In the present study, there were no significant differences in the ADC values between the craniopharyngiomas of the adamantinomatous and papillary types and the GCTs in the pure germinoma and non-pure germinoma groups. Craniopharyngiomas of the papillary type, however, are likely to have lower ADC values than those of the adamantinomatous type. Unfortunately, the subtype of GCT could not be analyzed in this study due to the small number of these lesions. Further studies may help to clarify the differences in ADC values between the pathological subtypes in craniopharyngiomas and GCTs and aid in the preoperative diagnosis of each subtype in the future.

Calcification is one of the characteristics in craniopharyngiomas. We tried to avoid the dense calcified lesions; however, the contamination of diffuse calcification might affect the ADC values of the craniopharyngiomas. Calcification itself has no signal on MR imaging, and the contamination of calcified lesions would make ADC values lower than original ADC values of the tumors. In other words, there is a possibility that the calcified craniopharyngiomas demonstrate lower ADC values than the non-calcified craniopharyngiomas. We compared the difference between the calcified and non-calcified tumors in our series, and we could not find a statistically significant difference between these two types of tumors. Our results implied that the influence of calcification could be reduced by carefully placing the ROIs on the non-calcified areas. In the end, the craniopharyngiomas with or without calcification showed markedly higher ADC values than the GCTs in the present study.

There are some limitations associated with our study. Sellar and parasellar lesions have high sensitivity to susceptibility artifacts due to nasal sinus air. Although we conducted PROPELLER MR imaging in patients with sellar and parasellar lesions, we could not completely exclude the susceptibility artifacts. In addition, craniopharyngiomas and GCTs include small cysts, microscopic calcification, and areas of microbleeding, all of which may affect ADC measurements. We compared CT and T1- and T2-weighted images on MR and carefully placed the ROIs in order to prevent the partial volume effect. Furthermore, 12 patients were excluded from the present study, because the effects of the cysts could not be avoided. Hence, our results should be fairly evaluated based on the effects of small cysts, calcification, and microbleeding, although we made every effort to remove these effects.

Conclusions

In this study, the craniopharyngiomas demonstrated lower intensity on DWI and higher ADC values at b = 1000 and b = 4000 s/mm2 on 3T MR imaging in comparison with the GCTs, especially with ADCMIN at b = 4000 s/mm2. DWI and the ADC values may help clinicians to differentiate between craniopharyngiomas and GCTs preoperatively.