Background: Silicone gel breast implants may silently rupture without detection. This has been the main reason for magnetic resonance imaging (MRI) of the augmented or reconstructed breast. The aim of the present study was to investigate the accuracy of MRI for implant rupture. Methods: Fifty consecutive patients with 85 silicone gel implants were included in the study. The mean age of the patients was 51 (range 21–72) years, with a mean duration of implantation of 3.8 (range 1–28) years. All patients underwent clinical examination and breast MRI. Intraoperative implant rupture was diagnosed by the operating surgeon. Results: Nineteen of the 50 patients suffered from clinical symptoms. An implant rupture was diagnosed by MRI in 22 of 85 implants (26%). In seven of 17 removed implants (41%), the intraoperative diagnosis corresponded with the positive MRI result. However, only 57% of these patients were symptomatic. Ultrasound imaging of the harvested implants showed signs of interrupted inner layers of the implant despite integrity of the outer shell. By microsurgical separation of the different layers of the implant shell, we were able to reproduce this phenomenon and to produce signs of implant rupture on MRI. Conclusion: Our results show that rupture of only the inner layers of the implant shell with integrity of the outer shell leads to a misdiagnosis on MRI. Correlation with clinical symptoms and the specific wishes of the patient should guide the indication for implant removal.