Title and Logo

First published as:
Rinckside –
Mapping the biological world
19 November 2017.


Rinckside
ISSN 2364-3889

Rinck PA.
Mapping the biological world.
Rinckside 2017; 28,7: 13-15.
Read the Print Edition (PDF)



Mapping the biological world

elaxation and relaxation constants is a rather complicated topic, both to explain and to under­stand. There are two main relaxation constants important for MRI: T1 and T2. T2* which is also often mentioned in this context is not a time constant, it is a capricious global parame­ter representing a fluctuant time or time range.


An excursion into scientific history

This is not the place to go into the science of relaxation; a textbook is better suited for this [1]. Here I will just tell a bit about the history and background of relaxation times in biomedicine.

It all began in 1953 with Eric Odeblad. He was the first to describe relaxation times in bio­logical systems. His first paper on the topic was enti­tled “Some preliminary observations on the proton magnetic resonance in biological samples” and pub­lished in Acta Radiologica Stockholm in early 1955 [2].

Odeblad had found that different tissues had dis­tinct relaxation times, most likely due to water con­tent but also to different bindings to lipids. Soon others joined in the new research field: Studies in blood, plasma and red blood cells, followed by T1- and T2-measurements of living frog muscle, the re­laxation of water in living animals and in the arms of living humans.

Research groups in Brooklyn and in Baltimore got involved in the early 1970s. They measured relax­ation times of ex­ci­sed normal and cancerous rat tis­sue, and the leader of the Brooklyn group stated that tumorous tissue had lon­ger relaxation times than normal tissue and promoted these findings as the ul­timate technology to screen for can­cer [3]. However, already some months later the Baltimore group stated that independent verification on the same NMR instru­ment could not be provided; the results were not reproducible [4].

Later, the New York Times pointed out major discrepan­cies between what was claimed by the researchers from Brooklyn and what was actu­ally accomplished, "discrepancies sufficient to make the author [Raymond Damadian] appear a fool if not a fraud." [5]

The summary of a paper published in 1975 – 42 years ago – by the group of Donald Hollis stated [6]:

“The direct use of NMR T1 measurements for can­cer diagnosis is clearly not feasible because of the lack of specificity … classification of tumors in this manner does not seem realistic.”

Shortly afterwards clinical MR imaging arrived and relaxation time mea­surements were considered very important during its first years. All ma­chines were programmed to create true T1 and T2 images (T1- and T2-mapping), based on reliable and reproducible spin-echo (SE) and inversion-recovery (IR) sequences. After absolute T1 and T2 values had been used un­successfully by re­searchers, combinations of T1 and T2, histogram techniques, and sophisticated three di­mensional dis­play techniques of factor representa­tions were used. At that time, these approaches were called ‘electronic contrast agents’, today ‘finger­printing’ or ‘biomark­ers’.

However, soon it became clear that relaxation time values were not the claimed invaluable addition to diagnostics, and these applications were skipped in the early 1990s.

“A spin-echo sequence with 24 echoes (Carr-Purcell-Meiboom-Gill sequence) was evaluated to determine the usefulness of magnetic resonance (MR) in detecting and typing brain tumors. ... T2 values calculated from an eight-point fit, however, did not allow discrimination of different tumors, nor did they allow differentiation between tumor, inflammatory tissue, and demyelination.” [7]

It was the time when the Relaxation Times Blues arrived [8], and Ian Young, one of the leading and influential scien­tists in MRI summed up the trials and errors in a short history of MRI as follows:

“Sadly, the many attempts that were made to corre­late pathology and relaxation be­havior have yielded none of the precise numerical relationships that were hoped for in the early days of MRI, so that this line of investigation ... has now been abandoned.” [9]


A grant-creating perpetuum mobile?

It is rare that a method appears, disappears, and then re-appears again as is the case of tissue characterization based on relaxation time constants. Yet some years later these obsolete methods were dug out again, grants were given to answer questions which had been discarded 25 years earlier [10, 11]. New pulse sequences and algorithms were developed – researchers tried their luck again.

Still, there is no easily explainable causality nor any evidence of a straight connection between these numbers and a distinct pathology. There is no unique signature of distinct malignancies or other pathologies in tissue relaxation times, be it in ex vivo or in vivo measurements. Many people believe that numbers (or, more fashionable, data) are the truth but they do not understand how the num­bers were acquired and what they stand for. Nature doesn't care about numbers. Believing in such postulations many years after they have been dismissed is a sign of scientific naiveté.

What's wrong in relaxation time mapping and applications: the pre­condition and presumption that a difficult biological structure such as a tissue or tissue changes in the human body can be quantified and qualified with NMR proton relaxation parameters.

Quantity and quality are being confused; it's so easy counting something – which doesn't mean that one can classify or characterize with numbers what one counts. The components and chemical and electrical pro­cesses in a tiny volume element, no matter how small it is, are far too complex and fickle to be expressed in bare figures. More so, on closer inspection, “objective" procedures, “objectively" defined range values as well as "objective" quality indicators for measurements often prove to be biased and interest-driven. There is no precise numerical fingerprint­ing based on relaxation constants in biomedicine.

It is helpful to once look into a microscope and to see how complex and complicated tissue structures are, both in normal and in pathological tissues – and in not-normal, but not (yet) pathological tissues.

In the end, it is not necessarily the errors or procedural “confounders” connected to the most elaborate and sophisti­cated data ac­quisition that make typing of normal and pathologi­cal tissues or grading of diseases impossible – but rather the complexity of tissue com­position and the over­lapping of relaxation time val­ues of heteroge­neous volume elements examined and processed into a sin­gle number or number range.

Nowadays lessons are rediscovered that became clear 25 years ago … and finally admitted, though diplomatically beating around the bush:

“In conclusion, our question, whether native T1 mapping in cardiac MR imaging can differentiate between healthy and diffuse diseased myocardium, must be answered with ‘yes’ and ‘no’, since the native myocardial T1 relaxation time allows discriminating between groups of patients with certain diffuse myocardial pathologies and a group of healthy individuals, but does not allow differencing between healthy and diffuse dis­eased myocardium in individual subjects.” [12]

Researchers also came to realize that novel methods for faster data acquisition deliver crude estimations but not accurate data. The higher the magnetic field, the larger seems to be the spread of T1 and T2 relaxation time estimations.

“A vast extent of methods and sequences has been developed to calculate the T1 and T2 relaxation times of different tissues in diverse centers. Surprisingly, a wide range of values has been reported for similar tissues (e.g. T1 of white matter from 699 to 1735 ms and T2 of fat from 41 to 371 ms), and the true values that represent each specific tissue are still unclear, which have deterred their common use in clinical diagnostic imaging.” [13]


Exceptions from the rule

Few iso­lated cases allow tissue discrimination based on re­laxation time alterations, but they are the excep­tion. One needs massive changes of relaxation time con­stants, as well as large homogeneously altered volumes to be able to use such data for diagnostic purposes. The data you get is not fake, it is not nec­essarily false, no, worse: it's half-true.

Does this mean that relaxation time maps cannot be used at all? Here are some insights into my own experiences:

We started creating maps of relaxation constants and proton density as well as derivatives of these maps, called “synthetic images”, in the early 1980s and pre­sent­ed the idea of synthetic MR images and si­mu­lat­ing entire MR exams in the early 1980s at a con­fe­ren­ce in the United States. In 1994 we finally published the image simulation software MR Image Expert for teaching and research purposes. More than 12,000 copies of MR Image Expert were licensed since then.

The simulations were based on the three main contrast parameters in MRI: T1, T2, and proton density acquired with time-consuming, but precise data acquisitions and exact calculations – with “clean” basic pulse sequences: inversion recovery and spin echo. For a reliable T1 determination one needs between 15 and 30 IR measurements, for T2 we usually used 24 echoes of a SE echo train. They allowed the creation of outstand­ingly good simulations of MR images – but still simulations.

In general, from a scientific point of view, MR imaging is a crude and not very exact technology. Thus, in most cases, relaxation time mapping and derivatives of it – such as synthetic images – cannot be used to quantify exact tissue data (e.g., relaxation constants or proton density in tis­sues) since the calculated or estimated relaxation constants and proton density values are unreliable – and impracticable in diagnostic routine; they are not accurate and not conclusive.

The only way to exploit relaxation time values would be situations when the values change drastically under specific physiological or pathological circumstances. This can be the case before and after the application of an MR contrast agent. There are uses for such rough estimations.

An area of application of relaxation times measurements might be the follow-up of massive T1 changes after the injection of a targeted contrast agent, such as Mn-DPDP and the comparison of plain and contrast-enhanced tissue, e.g., in heart diseases. Here imprecise measurements might be of diagnostic value.

However, such indications are limited because increasingly different and simpler MR techniques exist that may lead to the wanted result.

In one of the next columns I will try to discuss the non-scientific and non-medical reasons why these measurements returned and why they will stay with us for some time.


spaceholder blue

References

1. Rinck PA. Magnetic Resonance in Medicine. A Critical Introduction. The Basic Textbook of the EMRF. 12th edition, 2018 • to be published (details on www.magnetic-resonance.org).
2. Odeblad E, Lindström G. Some preliminary observations on the proton magnetic resonance in biological samples. Acta Radiol 1955; 43: 469-476.
3. Damadian RV. Tumor detection by nuclear magnetic resonance. Science 1971; 171: 1151-1153.
4. Hollis DP, Saryan LA, Morris HP. A nuclear magnetic resonance study of water in two Morris hepatomas. Johns Hopkins Med J. 1972; 131(6): 441-444.
5. Fjermedal G. Inside Dr. Damadian's magnet. The New York Times, 9 February 1986. www.nytimes.com/1986/02/09/books/inside-dr-damadian-s-magnet.html
6. Eggleston JC, Saryan LA, Hollis DP. Nuclear magnetic resonance investigations of human neoplastic and abnormal nonneoplastic tissues. Cancer Res. 1975; 35: 1326-1332.
7. Rinck PA, Meindl S, Higer HP, Bieler EU, Pfannenstiel P. Brain tumors: detection and typing by use of CPMG sequences and in vivo T2 measurements. Radiology. 1985; 157: 103-106.
8. Rinck PA. Relaxation times blues. Rinckside 1991; 2,1: 5-7.
9. Young I. Significant events in the development of MRI. J Magn Reson Imag JMRI 2004; 20: 183-186.
10. Rinck PA. Relaxing times for cardiologists. Rinckside 2015; 26,2: 3-5.
11. Rinck PA. MR fingerprinting returns to radiology – and hopefully disappears again. Rinckside 2015; 26,5: 13-14.
12. Goebel J, Seifert I, Nensa F, et al. Can native T1 mapping differentiate between healthy and diffuse diseased myocardium in clinical routine cardiac MR imaging? PLOS ONE | DOI:10.1371/journal.pone.0155591. 24 May 2016.
13. Bojorquez JZ, Bricq S, Acquitter C, Brunotte F, Walker PM, Lalande A. What are normal relaxation times of tissues at 3 T? Magn Reson Imaging. 2017 Jan;35:69-80. doi: 10.1016/j.mri.2016.08.021. Epub 2016 Sep 2.

spaceholder blue003 spaceholder blue003

VirtualCampus

00-f1

00-f2

00-f3

00-f4

00-f5

00-f1

00-f2

00-f3

00-f4

00-f5

00-f1

00-f2

00-f3

00-f4

00-f5

00-f1

00-f2

00-f3

00-f4

00-f5