Study Design

Data for this observational study were acquired prospectively and analyzed retrospectively. All data were provided by the NASA Lifetime Surveillance of Astronaut Health Program. The study was approved by the institutional review boards at the NASA Johnson Space Center and the Medical University of South Carolina. All participants provided written informed consent for the use and publication of their data. NASA reviewed the manuscript and the videos, which preserve the astronauts’ anonymity and are in compliance with the privacy standards of the NASA Astronaut Office.

Participants

Figure 1. Figure 1. Study Design.

A total of 35 astronauts underwent MRI of the brain before and after spaceflight. One astronaut underwent preflight and postflight imaging on two different MRI scanners and was excluded from the study. The cohort described here consisted of 34 astronauts, 18 of whom participated in long-duration missions (long-duration group) on the ISS and 16 of whom participated in short-duration missions (short-duration group) in space shuttle flights. Preflight MRIs were obtained at a mean of 287.5 days (range, 18 to 627) before launch. Postflight imaging occurred at a mean of 6.7 days (range, 1 to 20) after spaceflight. A total of 31 participants had traveled on previous missions and therefore had been exposed to microgravity before undergoing their preflight MRIs. The study design is summarized in Figure 1.

Imaging Protocols and Interpretation of Images

MRIs were performed at 3 Tesla, without the administration of contrast material, on a Philips Intera system (Philips Medical Systems) in 13 astronauts and on a Siemens Verio system (Siemens Healthcare) in 21 astronauts. The Philips imaging protocol included T 2 -weighted, multiplanar imaging (slice thickness, 3 to 4 mm), and the Siemens protocol included three-dimensional T 1 -weighted and T 2 -weighted imaging, which allowed for reconstruction of the slices in any plane (slice thickness, ≤1 mm).

On evaluation of T 2 -weighted imaging, the main prespecified analysis concerned change in the dimension of the central sulcus — whether it had widened, was unchanged, or had narrowed after spaceflight. Secondary features that were not prespecified but were considered areas of interest included change in the dimensions of the calcarine sulcus and supravermian cistern, the width of the third ventricle as measured on axial images, and the position of the cerebellar tonsils relative to the basion–opisthion line as measured on paired sagittal imaging. Measurements of the width of the third ventricle and the position of the cerebellar tonsils were made with the use of the measuring tool in the image-display software (Agfa Healthcare). All T 2 -weighted images were evaluated by two neuroradiologists who were unaware of flight durations and the order of presentation of preflight and postflight imaging. Masking of the images was facilitated by presenting MRIs from short- or long-duration flights in random sequence and presenting preflight and postflight scans for each participant in random sequence. If there was an lack of concordance between readers, the readers were required to reach consensus for the final reading.7 The quantitative measures (width of the third ventricle and position of the cerebellar tonsil) are presented as the mean (±SD) of the readers’ measurements.

Cine clips were used for additional analysis of a subgroup of study participants. These clips were generated from pairs of high-resolution, three-dimensional, T 1 -weighted sequences that were available for the 21 astronauts whose MRI studies were obtained in accordance with the Siemens protocol. The necessary postflight three-dimensional sequences were not performed for 1 astronaut in this group and the sequences for 2 others were degraded by motion. Therefore, cine clips were available for analysis from 18 astronauts (12 long-duration flights and 6 short-duration flights). The paired cine clips were compared by aligning an individual astronaut’s postflight MRI with the preflight MRI, allowing the reader to toggle between the two images in identical planes to facilitate a qualitative impression of change in anatomical relationships (for further details on these methods, see the Supplementary Appendix, available with the full text of this article at NEJM.org). To avoid confounding factors in the interpretation of anatomical changes, including differences in participant position within the scanner and differences in alignment between pairs of images, the outer table of the skull was used as a fixed reference (unlike weight-bearing bone, calvarial volume does not change in microgravity8). The cine clips were reviewed by two neuroradiologists to assess the following prespecified features: presence or absence of upward shift of the brain and narrowing of CSF spaces at the vertex. Other features that were evaluated but were not prespecified were cerebral aqueduct rotation (defined as any displacement or distortion of the aqueduct between preflight and postflight MRIs), uplifting of the optic chiasm, and stretching of the pituitary stalk.

Volumetric Ventricular Measurements

The percentage change in the total volume of the ventricular system (lateral, third, and fourth ventricles) between preflight and postflight was quantified in the 18 participants for whom high-resolution, three-dimensional, T 1 -weighted sequences were obtained. These sequences were analyzed with the use of the open-access Functional Magnetic Resonance Imaging of the Brain (FMRIB) Software Library, available at http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/. (For details, see the Supplementary Appendix.) The process of analyzing the sequences incorporated the following automated steps: brain extraction (removal of the skull and extracranial soft tissues), registration, and segmentation to obtain ventricular volumes.

Clinical Evaluation

All astronauts underwent ophthalmologic examination on return from space. Lumbar puncture was performed in those with optic-disk edema in order to measure opening pressure.

Statistical Analysis

We conducted Mann–Whitney U tests for independent samples (two-sided exact P value) to compare the two groups of astronauts on the basis of age, previous spaceflight experience, time in space, and time between preflight and postflight MRI examinations. Fisher’s exact tests were used to test the presence or absence of narrowing in specific CSF spaces in the short-duration and long-duration groups (two-sided exact P value, without corrections for multiple comparisons) and to evaluate the cine clips for the two groups for the presence or absence of brain movement. Independent sample Student’s t-tests were used to evaluate differences in the two groups of astronauts on the basis of preflight-to-postflight changes in the width of the third ventricle, position of the cerebellar tonsil, and total volume of the ventricular system. Statistical analyses were conducted with the use of SPSS software, version 23 (IBM).