def get_adenoma_histology(pm: PathManager):
"""
Uses JarManager to identify histology
:param pm: PathManager
:return:
"""
tub = pm.get_histology(Histology.TUBULAR)
tbv = pm.get_histology(Histology.TUBULOVILLOUS)
vil = pm.get_histology(Histology.VILLOUS)
return tub[0], tbv[0], vil[0]
def has_large_adenoma_precise(pm: PathManager, cm: CspyManager, min_size=10):
"""
Adds option of 'maybe'
WARNING: Logic will break down if >= 3 jars with the same location
:param pm:
:param cm:
:param min_size:
:return: 1=large adenoma, 0=no large adenoma, 9=maybe large adenoma
"""
if len(list(pm.get_locations_with_large_adenoma())) > 0:
logger.info('Found size and adenoma-ness in pathology')
return 1 # early exit: we found size and adenoma-ness
path_adenoma = list(pm.get_locations_with_unknown_adenoma_size())
# if not path_adenoma:
# return 0 # early exit: no adenomas
path_small = list(pm.get_locations_with_adenoma_size(max_size=min_size))
# NOTE: small adenoma not reliable due to fragments
path_adenoma += path_small
cspy_large = list()
cspy_small = list()
for f in cm.get_findings():
if f.size >= min_size:
for loc in f.locations_or_none():
cspy_large.append(loc)
else:
for loc in f.locations_or_none():
cspy_small.append(loc)
logger.info(f'Adenoma locations: {[str(ss) for ss in path_adenoma]}')
logger.info(f'Large polyp locations: {cspy_large}')
if not cspy_large:
return 0 # no large polyp
has_maybe = False
for aden_loc in path_adenoma:
if aden_loc in cspy_large: # is large adenoma
if aden_loc in cspy_small: # might be small
## NOTE: small path is unreliable
# if aden_loc in path_small: # here's the small one
# return 1
# else:
# if path_adenoma.count(aden_loc) == 1:
# return 1
if path_adenoma.count(aden_loc) < cspy_large.count(
aden_loc) + cspy_small.count(aden_loc):
has_maybe = True
else:
return 1
else:
return 1
return 9 if has_maybe else 0
def has_dysplasia(pm: PathManager):
"""
Identify mentions of highgrade dysplasia
:param pm:
:return:
"""
return 1 if pm.has_dysplasia() else 0
def has_large_adenoma_broad(pm: PathManager, cm: CspyManager, min_size=10):
"""
Location has large polyp and an adenoma
:param pm:
:param cm:
:param min_size:
:return:
"""
s = set(pm.get_locations_with_adenoma())
s2 = set(pm.get_locations_with_size(min_size))
for f in cm.get_findings_of_size(min_size):
if not f.locations:
s2.add(None)
for loc in f.locations:
s2.add(loc)
logger.info(f'Adenoma locations: {[str(ss) for ss in s]}')
logger.info(f'Large polyp locations: {s2}')
return 1 if (
s & s2 # both contain same location
or s2 and None in s # unknown adenoma location and large polyp in cspy
or s and None in s2 # unknown large polyp location (cspy)
) else 0
def get_villous_histology(pm: PathManager,
location: Location = Location.ANY,
allow_maybe=False):
"""
Get villous or tubulovillous by requested location
:param allow_maybe: if False, only include when location is guaranteed
:param location: location to look for
:param pm:
:return:
"""
tbv = pm.get_histology(Histology.TUBULOVILLOUS, allow_maybe=allow_maybe)
vil = pm.get_histology(Histology.VILLOUS, allow_maybe=allow_maybe)
if location == Location.ANY:
return 1 if tbv[0] + vil[0] else 0
elif location == Location.PROXIMAL:
return 1 if tbv[1] + vil[1] else 0
elif location == Location.DISTAL:
return 1 if tbv[2] + vil[2] else 0
elif location == Location.RECTAL:
return 1 if tbv[3] + vil[3] else 0
elif location == Location.UNKNOWN:
return 1 if tbv[4] + vil[4] else 0
else:
raise ValueError(f'Unrecognized location: {location}')
def process_text(path_text='',
cspy_text='',
cspy_finding_version=FindingVersion.BROAD):
pm = PathManager(path_text)
cm = CspyManager(cspy_text, version=cspy_finding_version)
data = {}
if pm:
specs, specs_combined, specs_dict = PathManager.parse_jars(path_text)
tb, tbv, vl = get_adenoma_histology(pm)
# count
adenoma_cutoff, adenoma_status, adenoma_count = get_adenoma_count_advanced(
pm)
_, _, jar_adenoma_count = get_adenoma_count_advanced(pm,
jar_count=True)
# distal
aden_dist_cutoff, aden_dist_status, aden_dist_count = get_adenoma_distal(
pm)
_, _, jar_ad_cnt_dist = get_adenoma_distal(pm, jar_count=True)
# proximal
aden_prox_cutoff, aden_prox_status, aden_prox_count = get_adenoma_proximal(
pm)
_, _, jar_ad_cnt_prox = get_adenoma_proximal(pm, jar_count=True)
# rectal
aden_rect_cutoff, aden_rect_status, aden_rect_count = get_adenoma_rectal(
pm)
_, _, jar_ad_cnt_rect = get_adenoma_rectal(pm, jar_count=True)
# unk
aden_unk_cutoff, aden_unk_status, aden_unk_count = get_adenoma_unknown(
pm)
_, _, jar_ad_cnt_unk = get_adenoma_unknown(pm, jar_count=True)
data.update({
ADENOMA_STATUS:
get_adenoma_status(specs),
TUBULAR:
tb,
TUBULOVILLOUS:
bool(tbv),
VILLOUS:
bool(vl),
ANY_VILLOUS:
get_villous_histology(pm),
PROXIMAL_VILLOUS:
get_villous_histology(pm, Location.PROXIMAL),
DISTAL_VILLOUS:
get_villous_histology(pm, Location.DISTAL),
RECTAL_VILLOUS:
get_villous_histology(pm, Location.RECTAL),
UNKNOWN_VILLOUS:
get_villous_histology(pm, Location.UNKNOWN),
SIMPLE_HIGHGRADE_DYSPLASIA:
get_highgrade_dysplasia(specs),
HIGHGRADE_DYSPLASIA:
has_dysplasia(pm),
ADENOMA_COUNT:
get_adenoma_count(specs),
LARGE_ADENOMA:
has_large_adenoma(pm, cm, version=cspy_finding_version),
ADENOMA_COUNT_ADV:
adenoma_count,
JAR_ADENOMA_COUNT_ADV:
jar_adenoma_count,
ADENOMA_STATUS_ADV:
adenoma_status,
ADENOMA_DISTAL:
aden_dist_status,
ADENOMA_DISTAL_COUNT:
aden_dist_count,
JAR_ADENOMA_DISTAL_COUNT:
jar_ad_cnt_dist,
ADENOMA_PROXIMAL:
aden_prox_status,
ADENOMA_PROXIMAL_COUNT:
aden_prox_count,
JAR_ADENOMA_PROXIMAL_COUNT:
jar_ad_cnt_prox,
ADENOMA_RECTAL:
aden_rect_status,
ADENOMA_RECTAL_COUNT:
aden_rect_count,
JAR_ADENOMA_RECTAL_COUNT:
jar_ad_cnt_rect,
ADENOMA_UNKNOWN:
aden_unk_status,
ADENOMA_UNKNOWN_COUNT:
aden_unk_count,
JAR_ADENOMA_UNKNOWN_COUNT:
jar_ad_cnt_unk,
JAR_SESSILE_SERRATED_ADENOMA_COUNT:
get_sessile_serrated_adenoma(pm, jar_count=True),
CARCINOMA_COUNT:
get_carcinomas(pm, jar_count=True),
CARCINOMA_MAYBE_COUNT:
get_carcinomas_maybe(pm, jar_count=True),
CARCINOMA_POSSIBLE_COUNT:
get_carcinomas_possible(pm, jar_count=True),
CARCINOMA_IN_SITU_COUNT:
get_carcinomas_in_situ(pm, jar_count=True),
CARCINOMA_IN_SITU_MAYBE_COUNT:
get_carcinomas_in_situ_maybe(pm, jar_count=True),
CARCINOMA_IN_SITU_POSSIBLE_COUNT:
get_carcinomas_in_situ_possible(pm, jar_count=True),
})
if cm:
data.update({
INDICATION: cm.indication,
NUM_POLYPS: cm.num_polyps,
BOWEL_PREP: cm.prep,
EXTENT: cm.extent,
})
# split maybe counters into two separate columns
data.update(split_maybe_counters(data))
return data
def get_carcinomas_in_situ_possible(pm: PathManager, jar_count=True):
return pm.get_carcinoma_in_situ_maybe_count(jar_count=jar_count)
def get_carcinomas_in_situ_maybe(pm: PathManager, jar_count=True):
return pm.get_carcinoma_in_situ_maybe_count(jar_count=jar_count,
probable_only=True)
def get_carcinomas_in_situ(pm: PathManager, jar_count=True):
return pm.get_carcinoma_in_situ_count(jar_count=jar_count)
def get_carcinomas(pm: PathManager, jar_count=True):
return pm.get_carcinoma_count(jar_count=jar_count)
def get_sessile_serrated_adenoma(pm: PathManager, jar_count=True):
return pm.get_sessile_serrated_count(jar_count=jar_count)