Purpose: To compare the femoral and tibial components rotational alignment in total knee arthroplasty (TKA) performed either with conventional or with patient-specific instrumentation. Methods: Forty-five patients underwent primary TKA and were prospectively randomized into two groups: 22 patients into the conventional instrumentation group (group A) and 23 patients into the Signature™ patient-specific instrumentation group (group B). All patients underwent computed tomography of the operated knee in the first week after surgery to measure the components rotation. Results: The femoral component rotation was 0.0° (-0.25, 1.0) in group A, and 0.0° (0.0, 1.0) in group B. The tibial component rotation was -16.0° (-18.5, 11.8) in group A, and -16.0° (-19.0, -14.0) in group B. There were no significant differences between the two groups in tibial and femoral components rotation. The difference between the tibial component rotation and the neutral tibial rotation was similar in both groups [2.0° (-0.5, 6.3) in group A and 2.0° (-1.0, 4.0) in group B], but the dispersion around the median was different between the two groups. The amplitude of the difference between tibial rotation and neutral position was 27° (-13, 14) in group A and 9° (-3, 6) in group B. Conclusions: There is a smaller chance of internal malrotation of the tibial component with the Signature™ patient-specific instrumentation system, with less dispersion and amplitude of the tibial component rotation around the neutral position. Level of evidence: II.