Fig. 13From: PredictPTB: an interpretable preterm birth prediction model using attention-based recurrent neural networksExample for a patient where PredictPTB was able to learn rare complications as risk factors for preterm birth. This patient was diagnosed with twin-to-twin transfusion syndrome (TTTS), a rare disorder which affects 10−15% of monochorionic, diamniotic twin pregnancies [50]. This observation is in line with literature reporting that pregnancies with TTTS complication are at increased risk for PTB [51]Back to article page