In Magnetic Resonance Imaging (MRI) of the lung, the pulmonary space is characterized by internal magnetic field inhomogeneities which originate from the susceptibility difference between the paramagnetic gas in the alveoli and the diamagnetic parenchymal tissue. These internal magnetic field inhomogeneities typically present spatial modulations at the level of the alveolar length scale [1,2], which is much smaller than the typical voxel size that is used in MRI. The corresponding microscopic magnetic field gradients, also called “internal gradients” [3,4], have a profound influence on the transverse relaxation mechanism in the lung tissue. In proton MRI, in fact, diffusion of water molecules through the internal gradients of the lung enhances the loss of phase coherence of the transverse magnetization [3,5]. This phenomenon manifests itself in a shortening of the transverse relaxation time (T2).

The initial evidence for the influence of diffusion on the T2 of the lung tissue was the observation of its dependence on the measurement method [6,7]. It was found, indeed, that the T2 values of the lung which are measured with the single spin-echo (or Hahn-echo) pulse sequence are considerably shorter than those which are obtained with a multi-echo spin-echo sequence [1,7,8]. The difference in the T2 values which were obtained with the two measurement methods was larger in inflated lungs than in airless lungs [1,7], indicating that the effect is enhanced by inflation. Such observation suggested that both 1) spin-spin interactions and 2) diffusion through the internal gradients regulate the T2 of the lung. For this reason, the transverse relaxation rate (1/T2) which is measured in the lung tissue and in other microscopically heterogeneous media is usually expressed as [5]:1T2=1T2,0+1T2,diffwhere 1/T2,0 is the intrinsic transverse relaxation rate which is due to spin-spin interactions and 1/T2,diff is the additional transverse relaxation rate which is due to diffusion through the internal gradients [5].

The transverse relaxation enhancement effect is usually considered as a drawback in lung MRI, because it accelerates signal decay. However, it also offers the possibility to assess interesting properties of the lung tissue. The value of T2,diff, in fact, does not only depend on the already mentioned lung inflation, but also on the alveolar size, as demonstrated by Kurz et al. [8]. The alveolar size is a parameter of considerable importance in the lung, as it represents a potential indicator of microstructural injuries. These can be induced by several lung diseases such as emphysema, Chronic Obstructive Lung Disease (COPD) and Acute Lung Injury (ALI), as well as by injury by mechanical ventilation [[9], [10], [11], [12]]. An additional lung property that can contribute to the transverse relaxation enhancement effect in-vivo is perfusion, due to the intravoxel incoherent motion (IVIM) of blood through the alveolar capillaries [2,13], which mimics a diffusion process also known as pseudo-diffusion. Because the lung tissue presents a multi-compartmental composition, in which the blood in the pulmonary capillaries constitutes a relevant fraction of the total lung water [13], the influence of perfusion on T2,diff, through the pseudo-diffusion coefficient of blood D⁎ [19], is possible. Other parameters which can affect T2,diff are the susceptibility difference between lung tissue and air in the alveoli (Δχ) and the static magnetic field used for the measurement (B0).

In earlier studies, the parameter T2,diff has been measured in porous materials [14] by using the well-known Hahn-echo pulse sequence [3,14]. In these studies, the measurement of the signal amplitude at different echo times (TE) was combined with a Laplace inversion in order to determine the statistical distribution of the internal gradients [4,14]. This approach, however, presents significant drawbacks with regard to the quantification of T2,diff of the lung in-vivo. These are: 1) knowledge of T2,0a-priori, or from a separate measurement, is required to determine T2,diff; 2) for the acquisition of each image, relatively long acquisition times of the order of several minutes are required, which are not suitable for in-vivo imaging of the lung; 3) the conventional Hahn echo approach is highly sensitive to respiratory motion.

Another possible approach to determine T2,diff consists in the analysis of the dependence of T2 on the time distance between two consecutive refocusing pulses (inter-pulse time) in a multi-echo spin-echo pulse sequence (Carr-Purcell mechanism) [4,5,7,8]. The basic principle of the Carr-Purcell mechanism is that the signal attenuation which is due to diffusion through the internal gradients is reduced by decreasing the inter-pulse time. This technique, which has been used for ex-vivo studies of the lung [6,7], requires multiple quantifications of T2 with different values of the inter-pulse time, resulting in relatively long acquisition times and high sensitivity to physiological motion, thus making it incompatible with in-vivo studies.

In this work, a novel technique based on proton MRI is presented, which allows to quantify T2,diff of the lung in a single breath-hold of approximately 10 s duration and without the need for a-priori knowledge of T2,0. To this end, a Half-Fourier-Acquired Single-shot Turbo spin-Echo (HASTE) [15] is used in combination with a Hahn-echo preparation module [16]. The underlying idea to quantify T2,diff is to exploit the Carr-Purcell mechanism by acquiring two images with identical TE, but with a different number of refocusing pulses between excitation and signal acquisition. The ability of this pulse sequence to quantify the parameter T2,diff was first validated in phantom by using the conventional Hahn-echo approach as a reference. In-vivo experiments were then conducted to investigate the influence of lung inflation and perfusion, through the cardiac phase, on T2,diff. The reason behind the investigation of the influence of the cardiac phase on T2,diff is the already mentioned effect of blood microcirculation in the alveolar capillaries, through the pseudo-diffusion coefficient D* [13,19]. D* depends on the blood velocity and, consequently, on the cardiac phase.