Our objective was to examine symptom-level changes in the course in posttraumatic stress disorder (PTSD) across the deployment cycle among combat-exposed Marines, and to determine the degree to which combat exposure and post-deployment stressor exposure predicted PTSD symptom profile transitions. We examined PTSD symptoms in a cohort of U.S. Marines (N = 892) recruited for the Marine Resiliency Study (MRS). Marines deployed as one battalion infantry unit to Afghanistan in 2010 and were assessed pre-deployment and one, five, and eight months post-deployment. We employed latent transition analysis (LTA) to examine Marines' movement across PTSD symptom profiles, determined by latent class analysis (LCA). LCAs revealed a 3-class solution one month pre-deployment, a 4-class solution at five months post-deployment, and a 3-class solution at eight months post-deployment. LTA revealed notable movement between classes over time, which depended chiefly on pre-deployment symptom presentation. Marines who reported few pre-deployment symptoms either maintained these low levels or returned to low levels by eight months. Marines who reported a moderate number of symptoms at pre-deployment had variable outcomes; 50% had reductions by eight months, and those who reported numbing symptoms at five months post-deployment tended to report more symptoms at eight months. Marines who reported more PTSD symptoms prior to deployment retained more symptoms eight months post-deployment. Combat exposure and post-deployment stressor exposure predicted profile transitions. Examining transitions between latent class membership over time revealed prognostic information about Marines' eight-month PTSD outcomes. The extent of pre-deployment PTSD symptoms was particularly informative of likely PTSD outcomes.