Cancelable biometric template protection (BTP) schemes are proposed to overcome some security issues that traditional biometric templates know once compromised in particular the non-renewability and the user trackability. The main concern of the existing BTP algorithms is to ensure the non-invertibility of the transform function and maintain high separability statistics to assess the security of the transformed template so that the original biometric data can’t be recovered back. We show in this paper that even if a BTP algorithm fulfills the above requirements, it is still vulnerable. We demonstrate this claim by launching an attack against a representative example of a transformation-based fingerprint template protection algorithm. The proposed attack, namely Attack via Missed Record Synchronization (AMRS), exploits the miss of shared redundant information between multiple correlated protected templates to deduce some additional hidden information that helps to reverse the transform function and spoof the authentication system. The proposed algorithm also allows to reconstruct the original template in terms of minutiae. The experiments conducted on the FVC database show that the proposed attack can break down the security of some cancelable algorithms knowing only reduced number of transformed templates.