Transarticular invasion of primary bone tumors abutting the sacroiliac joint: an MRI study in 128 patients

Background: To investigate magnetic resonance imaging (MRI) manifestation of transarticular invasion of primary bone tumors abutting the sacroiliac joint. Methods: We conducted a retrospective analysis of MRI data in 128 patients from January 2004 to December 2015. The diagnosis confirmed by surgical findings and pathological examination. The tumors of 87 patients located in the ilium and 41 patients in the sacrum. Tumors were divided into high-grade (Group 1), low-grade (Group 2) malignant tumor group and intermediate tumor group (Group 3). Transarticular invasion routes included 3 types: across ligament (Type 1), across articular cartilage (Type 2) and across periarticular tissue (Type 3). Results : Transarticular invasion were observed in 33 patients (25.8%): Group 1 (22 patients) included 15 patients (51.7%) with conventional osteosarcoma and 7 patients (43.8%) with Ewing's sarcoma; Group 2(5 patients) included 4 patients (11.4%) with central chondrosarcoma and 1 patient (7.7%) with chordoma; Group 3(6 patients) included 4 patients (20.0%) with giant cell tumor of bone, 1 patient (16.7%) with chondroblastoma and 1 patient (11.1%) with aneurysmal bone cyst. The difference of transarticular invasion incidence was not significant statistically between primary iliac and sacral tumors ( P > 0.05). The differences were significant statistically between group 2 and group 1, group 3 and group 1, respectively ( P < 0.01). Significant differences were observed among different invasion types ( P < 0.01): type 1(31 times), type 2(15 times) and type 3(5 times). Conclusions: MRI is sensitive and useful to evaluate transarticular invasion of primary bone tumors abutting the sacroiliac joint.

Due to the deep location of primary bone tumors abutting the sacroiliac joint, the tumors may invade the sacroiliac joint or even cross it at the time of diagnosis. Preoperative accurate diagnosis of transarticular invasion of the sacral or iliac tumor plays an important role in the design of surgical program [1] . Imaging studies of pelvic bone tumors invading the sacroiliac joint have been rarely reported in the literature [1][2][3][4] . Gadolinium-diethylenetriamine pentaacetic acid (Gd-DTPA) at a concentration of 0.5 mmol/L was intravenously administered at a dose of 0.2 ml/kg and at a rate of 1.0 ml/s for enhanced scans. The slice thickness was 5-8 mm, and the interlayer spacing was 0.5-0.8 mm.

Imaging evaluation
Sacral or iliac tumors with minimum length less than 2 cm from the tumor margin to the ipsilateral sacroiliac joint surface was defined as bone tumors abutting the sacroiliac joint [4] . The transarticular invasion was defined as sacral/iliac tumor invaded to the contralateral bone of the sacroiliac joint to cause bone destruction [1][2][3][4] . The

MRI diagnostic accuracy
Bone tumors abutting the sacroiliac joint can cause bone destruction and show abnormal MRI signal. The tumor tissue can invade into the sacroiliac joint or spread along the surrounding structures to the contralateral bone of the joint, resulting bone destruction and replacement with tumor tissue meanwhile. In this study, 33 of 128 (25.8%) patients presented with transarticular invasion (Table 1), which was 100% consistent with the surgical findings and pathological examination.

Comparison of transarticular invasion incidences in different locations
Of the 128 patients, 87 patients were diagnosed as iliac tumors, with 26 (29.9%) patients presenting with transarticular invasion; 41 patients had sacral tumors, with 7 (17.1%) patients presenting with transarticular invasion (Table 1). In general, the incidences between the iliac and sacral primary tumors were not different statistically (χ2 value = 2.39, P > 0.05). The difference of transarticular invasion incidence was not compared in Group 1 because their proportions were not compatible (43 iliac tumors and 2 sacral tumors). No statistically significant difference was observed in Group 2 26 iliac tumors and 22 sacral tumors, χ2 value = 0.04, P > 0.05 and Group 3 18 iliac tumors and 17 sacral tumors χ2 value = 0.14, P > 0.05 .

Comparison of the transarticular invasion incidences in different pathological grades
The overall difference of transarticular invasion incidences were significant statistically among different pathological grades (χ2 value = 19.84, P < 0.01). There were significant differences in Group 1 and Group2, Group 1 and Group 3 respectively (P <0.01) ( Table 2), but no difference between Group 2 and Group 3 statistically (χ2 value = 0.32, P > 0.05).

Comparison of the transarticular invasion incidences in different pathological types
The overall transarticular invasion incidence of malignant bone tumors (Group 1 + Group 2) was 29.0% (27/93). The transarticular invasion incidences were high in the patients with conventional osteosarcoma (51.7%) and those with Ewing's sarcoma (43.8%), but low in the patients with central chondrosarcoma (11.4%) and those with chordoma (7.7%) ( Table 1).The overall incidence of transarticular invasion was 17.1%

Comparison of transarticular invasion types
The details of transarticular invasion type were shown in Table 3. Type 1and type 1 + type 2 were common, whereas type 2 and type 3 were rare. Table 4 was the number of individual invasion type in the 33 patients. The overall transarticular invasion incidences were different among type 1, type 2 and type 3 (χ2 value = 41.74, P < 0.01). Further pairwise comparison showed statistic differences among the three types (χ2 values and corresponding P values were as follows: χ2 value = 18.37, P < 0.01; χ2 value = 41.31, P < 0.01; χ2 value = 7.17, P < 0.01). Transarticular invasion across the ligamentous portion was the most common type, followed by the articular cartilage type and finally, the periarticular tissue type. The incidences of type1/ type2/ type3 were not significantly different (P > 0.05) among the three groups respectively. In our study, the incidence of transarticular invasion was highest in Group 1(conventional osteosarcoma (51.7%) and Ewing's sarcoma (43.8%)), but was low in Group 2(central chondrosarcoma (11.4%) and chordoma (7.7%)). The incidence of osteosarcoma was similar to the results reported in the literature [2,4] . Nevertheless, the incidences of chondrosarcoma and Ewing's sarcoma were obviously different from that in the literature.

Discussion
Ozaki et al. reported in two different studies [2,4] that the transarticular invasion incidence of chondrosarcoma was 47.1% and 46.7% respectively, for Ewing's sarcoma was 4.3% and 8.7% respectively. We did not compared the results with other studies related to this because their sample sizes were too small [5][6][7] . These discrepancies may be attributed to differences in inclusion criteria. Firstly, Ozaki's study only included patients with the primary iliac tumor, whereas our study included patients with both primary iliac and sacral tumors. Secondly, minimum length of tumor margin away from the joint surface less than 2 cm was not an inclusion criteria in Ozaki's study [2] . Next, patients with highgrade chondrosarcoma accounted for 80% of all patients with chondrosarcoma in the studies by Ozaki [2,4] , whereas patients with low-grade chondrosarcoma (welldifferentiated central chondrosarcoma) were predominant in our study. For chondrosarcoma, the degree of malignancy may be a main reason for the transarticular invasion. The transarticular incidence of chordoma has not been reported in the literature and the incidence of it was only 7.7% in our study. This may be associated with its lowgrade malignancy and less invasive nature. In addition, chordoma is usually located at the midline of the lower sacrum [8] , which is distant from the sacroiliac joint relatively.
Of the 33 patients with tumor invasion across the sacroiliac joint, a few patients showed involvement of two or more invasion types. The statistical results showed differences in incidences among the 3 types: type 1 was the most common type, followed by type 2 and then, type 3. In addition, the incidences of different types were not associated with the malignant degree of tumors. Isolated invasion across the articular cartilage was very rare, so Type 2 was usually accompanied by type 1, indicating that the cartilage may prevent the joint to be invaded by the tumor to some degree. These findings were consistent with that in the previous literature [2,4] . The prevention mechanism of cartilage may be related to the following factors: ① There are no blood vessels in the articular cartilage. Thus, a direct anatomical channel and blood supply for tumor growth are lacking [8] . ② Cartilage cells can produce a substance to inhibit tumor angiogenesis [9] and collagenase activity [10][11][12] .

Conclusions
In summary, MRI can accurately diagnose transarticular invasion of primary bone tumors abutting the sacroiliac joint. The transarticular invasion incidence is associated with pathological type, benignity or malignancy, and the malignant degree of a tumor rather than its location (sacrum or ilium). The invasion type o is not related to the malignant degree of a tumor. Transarticular invasion across the ligamentous portion is quite common. The cartilage portion may be a barrier against tumor invasion, but as the tumor grows the cartilage can be destroyed and subsequently cause transarticular invasion.

Ethics Approval and Consent to Participate
The current study was approved by the Institutional Ethics Committee of the First Affiliated Hospital of Sun Yat-Sen University, and the need for signed informed consent was waived.

Consent for Publication
We have obtained consent to publish from all the participants.

Availability of data and materials
The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declared that they have no conflict of interest. No authors have received any funding from any institution, including personal relationships, interests, grants, employment, affiliations, patents, inventions, honoraria, consultancies, royalties, stock options/ownership, or expert testimony for the last 12 months.