Digital Signal Processing Reference
In-Depth Information
intracranial mass lesion is accurately made in 3090% of cases, depending on
tumor type. However, a biopsy is still considered the gold standard. Further
more, Howe and Opstad [207] point out that brain biopsy has a 1.7% mortality
rate, and note that in a study of 550 patients undergoing stereotactic biopsy:
8% had abscesses or inflammatory processes, 2.2% had other lesions, 3.4% of
the biopsies were nondiagnostic, and 8% of the patients had complications.
Thus, “a noninvasive and accurate prediction of lesion type would reduce
unnecessary surgical biopsy procedures for noncancerous lesions and for less
accessible tumors that would be treated by radio or chemotherapy rather
than surgical resection” (p.123).
8.1.2
Primary diagnosis of brain tumors by MRS & MRSI
The combination of anatomic plus molecular imaging via MRI, MRS and
MRSI is particularly promising for primary diagnosis of brain tumors. Brandao
and Domingues [211] note that MRI yields excellent spatial and contrast res
olution to assess structure. However, the need to assess more than the purely
anatomic aspects, such as biochemistry and tissue physiology, required the
development of functional techniques. “As a noninvasive method providing
metabolic information about the brain, MRS enables tissue characterization
at a biochemical level surpassing that of conventional magnetic resonance
imaging. MRS is also able to detect abnormalities that are invisible to MRI,
because metabolic abnormalities often precede structural changes” (p. 2).
8.1.2.1
Total choline
The hallmark for identifying neoplastic lesions using proton MRS has been
increased levels of choline; these are associated with high cell membrane
turnover and high cell density, as occurs with proliferation of brain tumor
cells. Sijens and Oudkerk [219] report a significantly higher mean choline
peak area in brain tumors (18 gliomas and 5 metastatic lesions) versus un
involved brain tissue. Similarly, Utriainen et al. [220] found significantly
elevated choline in twelve patients with suspected brain tumors compared to
MRS recordings from eleven healthy volunteers.
However, choline can be low in very small brain tumors and with necrosis
as in, e.g., glioblastoma multiforme. Pediatric brain tumors can also show
low choline [211, 221]. Low choline has also been reported in mixed tissues
that contained some brain tumor [222]. Nagar et al. [223] reported that
choline levels in intracranial tumors normalized to uninvolved brain tissue in
the same twenty patients were significantly higher than in fifteen patients with
nonneoplastic brain processes. However, in two of the fifteen patients with
nonneoplastic brain lesions the normalized choline levels were also elevated
(false positive rate 13%). High choline in a noncancerous lesion reportedly
also led to a false positive for malignancy in one of five children undergoing
MRS for workup of brainstem lesions [224].
Search WWH ::
Custom Search