Rectal Carcinoma: MRI with Histologic Correlation Before and After Chemoradiation Therapy
Steven D. Allen1,2,
Anwar R. Padhani2,
Andrzej S. Dzik-Jurasz3 and
Robert Glynne-Jones4
1 Department of Imaging, Royal Marsden Hospital, Downs Rd., Sutton, Surrey,
United Kingdom SM2 5PT.
2 Department of Imaging, Mount Vernon Cancer Centre, Northwood, Middlesex,
United Kingdom HA6 2RN.
3 Imaging Research, EPIX Pharmaceuticals, Inc., Cambridge, MA.
4 Department of Clinical Oncology, Mount Vernon Cancer Centre, Northwood,
Middlesex, United Kingdom HA6 2RN.
View larger version (154K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1 —61-year-old man with rectal cancer. Pretreatment axial T2-weighted
image through pelvis with small field of view shows large T3 rectal tumor with
extramural extension (black arrow) at 3-o'clock position. Tumor
extends into surrounding mesorectal fat but not to mesorectal fascia. Clear
circumferential resection margin (white arrow) of 4 mm is
evident.
|
|
View larger version (158K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2 —63-year-old woman with rectal cancer. Coronal T2-weighted image
through pelvis with small field of view shows bulky polypoid midrectal tumor
(black arrowheads) and enlarged right mesorectal lymph node
(white arrowhead). Tumor is well clear of outer margin of mesorectal
fascia (white arrows), which can be easily traced down to levator
muscle (black arrows).
|
|
View larger version (170K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3 —72-year-old man with rectal cancer. Axial T2-weighted image through
pelvis with small field of view shows annular tumor that appears confined by
rectal wall. Tumor contains heterogeneous intermediate and high signal
intensity (black arrows) in keeping with mucin production.
Low-signal-intensity mesorectal fascia (white arrow) outlines
mesorectum.
|
|
View larger version (138K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A —21-year-old woman with rectal cancer. Nodal downstaging independent
of poor tumor response to treatment. Sagittal T2-weighted image shows large
low rectal tumor (black arrow) and multiple enlarged mesorectal nodes
(white arrows) in posterior aspect.
|
|
View larger version (151K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B —21-year-old woman with rectal cancer. Nodal downstaging independent
of poor tumor response to treatment. Sagittal T2-weighted image obtained after
chemoradiotherapy shows large low rectal tumor (arrow) in A
has not changed in length despite treatment. Mesorectal nodal deposits are
much smaller. Three small tumor-free nodes were recovered at histologic
examination.
|
|
View larger version (138K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A —55-year-old man with rectal cancer. Good tumor response to
chemoradiation. Axial T2-weighted image of polypoidal rectal cancer shows
tumor interpreted as having few millimeters of extramural tumor (T3)
(white arrow). Borderline enlarged right mesorectal lymph node
(black arrow) is evident.
|
|
View larger version (159K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B —55-year-old man with rectal cancer. Good tumor response to
chemoradiation. Axial T2-weighted image at same level as A after
chemoradiotherapy shows excellent response to treatment, with rectal tumor no
longer visible. Low-signal-intensity muscle wall fibrosis (arrow) is
evident. No mesorectal nodes are present. Histologic findings confirmed
complete tumor response to treatment and posttreatment scarring.
|
|
View larger version (139K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6B —45-year-old man with rectal cancer. Apparent good response to
treatment. Sagittal T2-weighted image after chemoradiotherapy shows marked
reduction in tumor bulk. Postchemoradiation therapy sacral bone marrow fatty
atrophy (arrow) is evident. Histologic examination showed small
amount of T3 disease remained. Differentiating small-volume extramural tumor
from fibrosis after chemotherapy and radiation treatment is diagnostically
difficult and is common source of MRI inaccuracy in tumor staging.
|
|
View larger version (146K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7A —66-year-old man with rectal cancer. Mucinous tumor. Axial
T2-weighted image through pelvis shows extramural tumor (arrow) at 3
-o'clock position. Tumor is of mixed intermediate and relatively high signal
intensity, indicative of mucinous histologic characteristics.
|
|
View larger version (166K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7B —66-year-old man with rectal cancer. Mucinous tumor. Axial
T2-weighted image at same level as A. After chemoradiation therapy,
signal intensity within tumor has increased and is closer to uniform. Tumor
itself (arrow) has otherwise changed little in size or structure.
These appearances are misleading because no active tumor but only inactive
mucin lakes were present at subsequent histologic examination.
|
|
View larger version (8K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8A —Bland and Altman plots show accuracy of prediction of lesion size
and distance of tumor to circumferential resection margin (CRM) using
histology as gold standard. Mean difference (solid line) and 95%
limits of agreement (hashed lines) are indicated in each graph.
Scatter graphs at top of all graphs show spread in all tumors; graphs at
bottom show spread with exclusion of mucinous tumors. From these we can see
that lesion length (A and B) is not well predicted. For all
tumors (A and B, top), there is a tendency to overestimate
lesion length by 5.4 mm, with wide range of 35 mm to 46 mm. For the more
important parameter, distances of tumor to CRM (C and D),
predictions are better. For all tumors (C and D, top) mean
difference between MRI and histology is only 0.2 mm and range is also narrow
(8 to 5 mm).
|
|
View larger version (7K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8B —Bland and Altman plots show accuracy of prediction of lesion size
and distance of tumor to circumferential resection margin (CRM) using
histology as gold standard. Mean difference (solid line) and 95%
limits of agreement (hashed lines) are indicated in each graph.
Scatter graphs at top of all graphs show spread in all tumors; graphs at
bottom show spread with exclusion of mucinous tumors. From these we can see
that lesion length (A and B) is not well predicted. For all
tumors (A and B, top), there is a tendency to overestimate
lesion length by 5.4 mm, with wide range of 35 mm to 46 mm. For the more
important parameter, distances of tumor to CRM (C and D),
predictions are better. For all tumors (C and D, top) mean
difference between MRI and histology is only 0.2 mm and range is also narrow
(8 to 5 mm).
|
|
View larger version (7K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8C —Bland and Altman plots show accuracy of prediction of lesion size
and distance of tumor to circumferential resection margin (CRM) using
histology as gold standard. Mean difference (solid line) and 95%
limits of agreement (hashed lines) are indicated in each graph.
Scatter graphs at top of all graphs show spread in all tumors; graphs at
bottom show spread with exclusion of mucinous tumors. From these we can see
that lesion length (A and B) is not well predicted. For all
tumors (A and B, top), there is a tendency to overestimate
lesion length by 5.4 mm, with wide range of 35 mm to 46 mm. For the more
important parameter, distances of tumor to CRM (C and D),
predictions are better. For all tumors (C and D, top) mean
difference between MRI and histology is only 0.2 mm and range is also narrow
(8 to 5 mm).
|
|
View larger version (6K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8D —Bland and Altman plots show accuracy of prediction of lesion size
and distance of tumor to circumferential resection margin (CRM) using
histology as gold standard. Mean difference (solid line) and 95%
limits of agreement (hashed lines) are indicated in each graph.
Scatter graphs at top of all graphs show spread in all tumors; graphs at
bottom show spread with exclusion of mucinous tumors. From these we can see
that lesion length (A and B) is not well predicted. For all
tumors (A and B, top), there is a tendency to overestimate
lesion length by 5.4 mm, with wide range of 35 mm to 46 mm. For the more
important parameter, distances of tumor to CRM (C and D),
predictions are better. For all tumors (C and D, top) mean
difference between MRI and histology is only 0.2 mm and range is also narrow
(8 to 5 mm).
|
|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Copyright © 2007 by the American Roentgen Ray Society.
