Clinical
Histopathologic Staging in Laryngeal Cancer Harold R. C.
vs
Pillsbury, MD, John
A.
Kirchner,
\s=b\ One hundred fourteen serially sectioned, nonirradiated, laryngeal speci-
that were removed for cancer have been examined to determine the accuracy of preoperative staging. Characteristics of the primary lesion, which indicate invasion of the laryngeal framework, have been demonstrated for each region in the larynx. Pathologic findings, such as submucosal extension, growth into the preepiglottic space, infraglottic extension, and involvement of the laryngeal ventricle, have been correlated with clinical staging. The relationship between tumor size and accuracy of clinical staging has been evaluated. For multiregional lesions, the location that yielded the highest probability of metastasis was designated as the primary site. Our findings indicate inaccurate staging for 37% of glottic tumors, 38% of supraglottic tumors, 50% of transglottic tumors, and 13% of subglottic lesions. In 89% of the cases that were staged inaccurately, the error was one of underestimation. Depth of tumor invasion was the most difficult measurement to determine. mens
(Arch Otolaryngol 105:157-159, 1979) factor in most can¬ treatment of in is the accuracy of clini¬
important plan¬ Thening laryngeal patients
cer
cal
staging. Characteristics of the
for publication Aug 28, 1978. From the Section of Otolaryngology, Yale University School of Medicine, New Haven,
Accepted
Conn.
Reprint requests to Section of Otolaryngology, Yale University School of Medicine, 333 Cedar St, New Haven, CT 06510 (Dr Pillsbury).
MD
primary
lesion that indicate deep often difficult to assess. Specific locations, such as the anterior commissure, laryngeal ventricle, and subglottis, are hard to examine in the presence of adjacent carcinoma. In addition, there are areas within the larynx, such as the arytenoid carti¬ lage, where mucosal involvement by tumor may or may not be associated with invasion of underlying cartilage. Indirect and direct laryngoscopy, lar¬
invasion
are
yngeal tomography, contrast laryngography, and xeroradiography have all been used in evaluating tumor extent. Our objectives include a more accurate method of identifying those lesions that might recur locally or regionally and that might, therefore, be candi¬ dates for combined treatment modali¬ ties. MATERIAL AND METHODS
One hundred fourteen laryngeal cancers studied for size, location, and depth of invasion. Each larynx was removed by total laryngectomy, opened posteriorly in the midline, and photographed. After a small sample of tumor had been removed for the pathologist, all specimens were processed according to the technique developed by Tucker,1 and studied by serial sections that were made at 20-µ intervals. Our material from the time period was gathered between 1958 and 1976 and includes only those cases in which no radiotherapy was given. Clinical staging was updated accord¬ ing to the standards set by the Union were
Internationale Contre de Cancer in 19742 on the basis of the clinician's narrative description of tumor extent and degree of fixation at the time of initial staging, so that errors due to changes in the nomencla¬ ture of the clinical staging system would not affect the results of this study. Any tumor that extended more than 5 mm below the aryepiglottic fold along the medial wall of the pyriform sinus, or 5 mm below the interarytenoid space into the
postcricoid area, was assigned a hypopha¬ ryngeal classification and excluded from
this study. The system that was developed for histopathologic staging is modified from the one proposed by Norris et al' as follows:
P„—No tumor found
P,-Superficial carcinoma
in mucosa or submucosa confined to one anatomic site P2-Superficial carcinoma in mucosa or submucosa in more than one anatomic site P8—Evidence of deep invasion up to the limits of the larynx: Muscle invasion in glottic cases Involvement of lingual surface of epiglottis into vallecula Involvement of preepiglottic space not beyond thyrohyoid membrane Submucosal extension into medial wall of pyriform sinus P4-Beyond the limits of the larynx: Extension into the laryngeal framework Involvement of base of tongue Extension through thyrohyoid membrane Extension into wall of trachea
Downloaded From: http://archotol.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/29/2015
Table 1.—Incorrect
Staging
Correctly Staged
Subglottic (N 6) Glottic (N 32) Supraglottic (N 34) Transglottic (N 42)
87 64 63 50
=
=
=
Table Clinical
Laryngeal Cancer
Tumor, %
Location of Tumor =
Overstaged
Understaged 13 29 29 50
2.—Staging Errors* No. (%) of Errors
vs
Pathologic Staging T,;P, T; P3 T„ T,; P, Overstaged
1(1) (3) 36 (32) 5 (4) 46 (40) 4
Total
ly·
in 114 Cases of
'Forty-six of 114 cases were staged incorrect¬
Fig 2.—Sagittal section of growth (arrow) crossing anterior commissure in vertical direction and demonstrating thyroid inva¬ sion. Such lesions usually invade and destroy laryngeal framework and, there¬ fore, should be classified as T„. EP indi¬ cates epiglottis; HY, hyoid (hematoxylin1).
eosin,
whelming majority of staging errors—
(89%) consisted of underesti¬ mating tumor size (Table 2). Out of 32 cases of glottic lesions, there were 12 staging errors. Ten cases were understaged because of the following reasons: laryngeal frame¬ work was invaded in seven; infraglot41/46—
1 .—Growth involves base of epiglottis, but extends downward onto and below true cords at anterior commissure. Both cords were mobile so lesion was staged T,. Sections showed thyroid cartilage was destroyed by cancer at anterior commis¬ sure area, making lesion T,. Invasion of laryngeal framework has usually been found in other similar lesions that cross anterior commissure in vertical direction.
Fig
specimens were evaluated by investigators (H.R.C.P. and J.A.K.).
All
both
were
RESULTS
Forty-six (40%)
of the 114
tic extension occurred in two; and one was attached to the internal perichon¬ drium. In most of these understaged cases, the clinician failed to anticipate cartilage invasion. Two cases, which
cases
in
this series were staged incorrectly. Table 1 depicts the number of cases for each location in which errors were made. While a small percentage of tumors were overstaged, the over-
overstaged,
were verrucous car¬
cinomas that seemed to fix the vocal cord by their bulk. Pathologic exami¬ nation disclosed these lesions to be
superficial. Staging errors occurred in 13 of 34 cases of supraglottic tumors. Lesions of the suprahyoid epiglottis were easi-
stage accurately than those in infrahyoid portion. Extension of infrahyoid tumor into the preepiglot-
er
to
the
tic space and base of the tongue was particularly difficult to detect, which accounted for five of the ten understaged miscalculations. Three lesions on the laryngeal surface of the epiglottis, which extended below the anterior commissure, were not sus¬ pected of invading cartilage. There was one lesion with extension beyond the thyrohyoid membrane and one with unrecognized extension along the medial wall of the pyriform sinus. These were actually transglottic carci¬ nomas, and all showed invasion of the thyroid cartilage on serial sections (Fig 1). By contrast, none of the supraglottic tumors that remained above the anterior commissure in¬ vaded cartilage.4 There were three overstaged supraglottic lesions: two with no extension through the thyro¬ hyoid membrane, as suspected clini¬ cally, and one with no extension into the base of the tongue, also suspected
clinically. Transglottic cancers were the most difficult to stage correctly. There were staging errors in 21 of 42 cases. Unsuspected invasion of the laryngeal framework was discovered in 17 cases, nearly always in the ossified portions (thyroid ala and cricoid area). Size is a factor in invasion by transglottic cancer, since lesions under 2 cm did not invade the laryngeal framework, while more than 75% of the lesions that were larger than 3 cm did. For invasion of the thyroid ala to occur, the growth must cross the ventricle itself, and not simply extend down¬ ward behind it3 (77% for the former and 0% for the latter, in 34 and 11 cases, respectively). The following causes also occurred in understaged transglottic lesions: extralaryngeal spread via cricothyroid membrane (two cases); extensions into base of
tongue (one case); and extension to trachea (one case). Seven transglottic growths were originally considered supraglottic, but were found to extend below the ante¬ rior commissure at the petiole of the epiglottis. Four of these were discov¬ ered preoperatively and, therefore, do not represent staging errors. Six of
Downloaded From: http://archotol.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/29/2015
3.—Most transglottic lesions, measur¬ 3 cm or more by surface presentation, invade and destroy thyroid ala as shown (arrow). Since these extend beyond con¬ fines of larynx, these should be staged T4. indicates thyroid cartilage.
Fig ing
Fig 4—Coronal section of transglottic growth (arrow) demonstrates conus elasticus and illustrates tendency of cancer within membrane compartment to escape from larynx via cricothyroid membrane (Verhoeff's stain, x1).
these seven lesions demonstrated thy¬ roid cartilage invasion (Fig 2). It is difficult to perceive these as T4 tumors because of associated vocal cord mobility. Nevertheless, these escape the larynx by direct extension. Therefore, invasion of the thyroid cartilage at the anterior commissure is to be expected with a growth at the base of the epiglottis, which extends down onto or below the vocal cords. Fewer errors were made with staging subglottic cancers because all initially appeared with a fixed vocal cord. One subglottic growth that was staged as a T, was found, at examination of serial sections, to extend paratracheally; therefore, this growth became a P4. Glottic and supraglottic cancer that extended onto the arytenoid cartilage was studied for invasion into the cartilage itself, and for its relation to such clinical features as vocal cord fixation and extent of visible mucosal spread. The total numbers of such cases in each category were too small to allow for conclusions. However, our impression from these few cases, is that vocal cord fixation in the presence of clinical involvement of the arytenoid is usually associated with deep invasion of the thyroarytenoid muscle; it is this latter feature that is responsible for this fixation. Assessment of invasion of the laryngeal framework is difficult. Xeroradiography is often helpful, but laryngeal tomography is more reli¬ able. Lateral laryngography is useful in evaluating the anterior commissure in the presence of cancer at the base of the epiglottis. It may also reveal preepiglottic space invasion by show¬ ing posterior displacement of the epiglottis. Roentgenographic demon¬ stration of the ventricle is sometimes misleading: a bulky lesion of the superior surface of the vocal cord may mimic a ventricular or small, trans¬ glottic cancer. The ventricle is best examined at direct laryngoscopy by pushing aside the lower edge of the ventricular band or by pulling it upward on a hook. Submucosal tumor in the base of the tongue is difficult to evaluate without palpation under general anesthesia.
COMMENT
The 40% rate of error in staging resulted mainly from a failure to predict extension of cancer outside the confines of the larynx. Therefore, there should have been more T4 and fewer T3 lesions in the original clinical staging. However, these errors might not affect the choice of a surgical procedure in supraglottic cancer, where the preepiglottic space is removed, as completely with horizon¬ tal supraglottic partial laryngectomy, as with total laryngectomy. However, with invasion into the base of the tongue, understaging could be disas¬ trous.
transglottic lesions, failure to anticipate extension of cancer through the thyroid or cricoid carti¬ lages, or through the cricothyroid membrane, could lead to inadequate surgical resection (Fig 3 and 4). Simi¬ larly, understaging may result in poor control rates when radiotherapy is the sole modality. With
CONCLUSIONS Present staging systems fail to identify invasion and destruction of the laryngeal framework in a high percentage of cases. This feature accounts for most of the 40% error in predicting the growth patterns of laryngeal cancer. Since tumor volume
presents
a
greater therapeutic prob¬
lem to the radiotherapist than to the surgeon, unsuspected deep invasion of tumor may account for relatively poorer control rates by radiotherapy than by surgery for any particular
stage. References 1. Tucker GF Jr: A histological method for the study of the spread of carcinoma within the larynx. Ann Otol 70:910-921, 1961. 2. TMN Classification of Malignant Tumors, ed 2. Geneva, Union Internationale Contre de Cancer, 1974. 3. Norris CM, Kuo BF, Tucker GF Jr, et al: A correlation of clinical staging, pathological findings and five-year end results in surgically treated cancer of the larynx. Ann Otol Rhinol Laryngol 79:1033-1048, 1970. 4. Kirchner JA, Som ML: Clinical and histological observations on supraglottic cancer. Ann Otol Rhinol Laryngol 80:638-645, 1971. 5. Kirchner JA, Cornog JL Jr, Holmes RE: Transglottic cancer: Its growth and spread within the larynx. Arch Otolaryngol 99:247-251, 1974.
Downloaded From: http://archotol.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/29/2015
Histopathologic Staging in Laryngeal Cancer Harold R. C.
vs
Pillsbury, MD, John
A.
Kirchner,
\s=b\ One hundred fourteen serially sectioned, nonirradiated, laryngeal speci-
that were removed for cancer have been examined to determine the accuracy of preoperative staging. Characteristics of the primary lesion, which indicate invasion of the laryngeal framework, have been demonstrated for each region in the larynx. Pathologic findings, such as submucosal extension, growth into the preepiglottic space, infraglottic extension, and involvement of the laryngeal ventricle, have been correlated with clinical staging. The relationship between tumor size and accuracy of clinical staging has been evaluated. For multiregional lesions, the location that yielded the highest probability of metastasis was designated as the primary site. Our findings indicate inaccurate staging for 37% of glottic tumors, 38% of supraglottic tumors, 50% of transglottic tumors, and 13% of subglottic lesions. In 89% of the cases that were staged inaccurately, the error was one of underestimation. Depth of tumor invasion was the most difficult measurement to determine. mens
(Arch Otolaryngol 105:157-159, 1979) factor in most can¬ treatment of in is the accuracy of clini¬
important plan¬ Thening laryngeal patients
cer
cal
staging. Characteristics of the
for publication Aug 28, 1978. From the Section of Otolaryngology, Yale University School of Medicine, New Haven,
Accepted
Conn.
Reprint requests to Section of Otolaryngology, Yale University School of Medicine, 333 Cedar St, New Haven, CT 06510 (Dr Pillsbury).
MD
primary
lesion that indicate deep often difficult to assess. Specific locations, such as the anterior commissure, laryngeal ventricle, and subglottis, are hard to examine in the presence of adjacent carcinoma. In addition, there are areas within the larynx, such as the arytenoid carti¬ lage, where mucosal involvement by tumor may or may not be associated with invasion of underlying cartilage. Indirect and direct laryngoscopy, lar¬
invasion
are
yngeal tomography, contrast laryngography, and xeroradiography have all been used in evaluating tumor extent. Our objectives include a more accurate method of identifying those lesions that might recur locally or regionally and that might, therefore, be candi¬ dates for combined treatment modali¬ ties. MATERIAL AND METHODS
One hundred fourteen laryngeal cancers studied for size, location, and depth of invasion. Each larynx was removed by total laryngectomy, opened posteriorly in the midline, and photographed. After a small sample of tumor had been removed for the pathologist, all specimens were processed according to the technique developed by Tucker,1 and studied by serial sections that were made at 20-µ intervals. Our material from the time period was gathered between 1958 and 1976 and includes only those cases in which no radiotherapy was given. Clinical staging was updated accord¬ ing to the standards set by the Union were
Internationale Contre de Cancer in 19742 on the basis of the clinician's narrative description of tumor extent and degree of fixation at the time of initial staging, so that errors due to changes in the nomencla¬ ture of the clinical staging system would not affect the results of this study. Any tumor that extended more than 5 mm below the aryepiglottic fold along the medial wall of the pyriform sinus, or 5 mm below the interarytenoid space into the
postcricoid area, was assigned a hypopha¬ ryngeal classification and excluded from
this study. The system that was developed for histopathologic staging is modified from the one proposed by Norris et al' as follows:
P„—No tumor found
P,-Superficial carcinoma
in mucosa or submucosa confined to one anatomic site P2-Superficial carcinoma in mucosa or submucosa in more than one anatomic site P8—Evidence of deep invasion up to the limits of the larynx: Muscle invasion in glottic cases Involvement of lingual surface of epiglottis into vallecula Involvement of preepiglottic space not beyond thyrohyoid membrane Submucosal extension into medial wall of pyriform sinus P4-Beyond the limits of the larynx: Extension into the laryngeal framework Involvement of base of tongue Extension through thyrohyoid membrane Extension into wall of trachea
Downloaded From: http://archotol.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/29/2015
Table 1.—Incorrect
Staging
Correctly Staged
Subglottic (N 6) Glottic (N 32) Supraglottic (N 34) Transglottic (N 42)
87 64 63 50
=
=
=
Table Clinical
Laryngeal Cancer
Tumor, %
Location of Tumor =
Overstaged
Understaged 13 29 29 50
2.—Staging Errors* No. (%) of Errors
vs
Pathologic Staging T,;P, T; P3 T„ T,; P, Overstaged
1(1) (3) 36 (32) 5 (4) 46 (40) 4
Total
ly·
in 114 Cases of
'Forty-six of 114 cases were staged incorrect¬
Fig 2.—Sagittal section of growth (arrow) crossing anterior commissure in vertical direction and demonstrating thyroid inva¬ sion. Such lesions usually invade and destroy laryngeal framework and, there¬ fore, should be classified as T„. EP indi¬ cates epiglottis; HY, hyoid (hematoxylin1).
eosin,
whelming majority of staging errors—
(89%) consisted of underesti¬ mating tumor size (Table 2). Out of 32 cases of glottic lesions, there were 12 staging errors. Ten cases were understaged because of the following reasons: laryngeal frame¬ work was invaded in seven; infraglot41/46—
1 .—Growth involves base of epiglottis, but extends downward onto and below true cords at anterior commissure. Both cords were mobile so lesion was staged T,. Sections showed thyroid cartilage was destroyed by cancer at anterior commis¬ sure area, making lesion T,. Invasion of laryngeal framework has usually been found in other similar lesions that cross anterior commissure in vertical direction.
Fig
specimens were evaluated by investigators (H.R.C.P. and J.A.K.).
All
both
were
RESULTS
Forty-six (40%)
of the 114
tic extension occurred in two; and one was attached to the internal perichon¬ drium. In most of these understaged cases, the clinician failed to anticipate cartilage invasion. Two cases, which
cases
in
this series were staged incorrectly. Table 1 depicts the number of cases for each location in which errors were made. While a small percentage of tumors were overstaged, the over-
overstaged,
were verrucous car¬
cinomas that seemed to fix the vocal cord by their bulk. Pathologic exami¬ nation disclosed these lesions to be
superficial. Staging errors occurred in 13 of 34 cases of supraglottic tumors. Lesions of the suprahyoid epiglottis were easi-
stage accurately than those in infrahyoid portion. Extension of infrahyoid tumor into the preepiglot-
er
to
the
tic space and base of the tongue was particularly difficult to detect, which accounted for five of the ten understaged miscalculations. Three lesions on the laryngeal surface of the epiglottis, which extended below the anterior commissure, were not sus¬ pected of invading cartilage. There was one lesion with extension beyond the thyrohyoid membrane and one with unrecognized extension along the medial wall of the pyriform sinus. These were actually transglottic carci¬ nomas, and all showed invasion of the thyroid cartilage on serial sections (Fig 1). By contrast, none of the supraglottic tumors that remained above the anterior commissure in¬ vaded cartilage.4 There were three overstaged supraglottic lesions: two with no extension through the thyro¬ hyoid membrane, as suspected clini¬ cally, and one with no extension into the base of the tongue, also suspected
clinically. Transglottic cancers were the most difficult to stage correctly. There were staging errors in 21 of 42 cases. Unsuspected invasion of the laryngeal framework was discovered in 17 cases, nearly always in the ossified portions (thyroid ala and cricoid area). Size is a factor in invasion by transglottic cancer, since lesions under 2 cm did not invade the laryngeal framework, while more than 75% of the lesions that were larger than 3 cm did. For invasion of the thyroid ala to occur, the growth must cross the ventricle itself, and not simply extend down¬ ward behind it3 (77% for the former and 0% for the latter, in 34 and 11 cases, respectively). The following causes also occurred in understaged transglottic lesions: extralaryngeal spread via cricothyroid membrane (two cases); extensions into base of
tongue (one case); and extension to trachea (one case). Seven transglottic growths were originally considered supraglottic, but were found to extend below the ante¬ rior commissure at the petiole of the epiglottis. Four of these were discov¬ ered preoperatively and, therefore, do not represent staging errors. Six of
Downloaded From: http://archotol.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/29/2015
3.—Most transglottic lesions, measur¬ 3 cm or more by surface presentation, invade and destroy thyroid ala as shown (arrow). Since these extend beyond con¬ fines of larynx, these should be staged T4. indicates thyroid cartilage.
Fig ing
Fig 4—Coronal section of transglottic growth (arrow) demonstrates conus elasticus and illustrates tendency of cancer within membrane compartment to escape from larynx via cricothyroid membrane (Verhoeff's stain, x1).
these seven lesions demonstrated thy¬ roid cartilage invasion (Fig 2). It is difficult to perceive these as T4 tumors because of associated vocal cord mobility. Nevertheless, these escape the larynx by direct extension. Therefore, invasion of the thyroid cartilage at the anterior commissure is to be expected with a growth at the base of the epiglottis, which extends down onto or below the vocal cords. Fewer errors were made with staging subglottic cancers because all initially appeared with a fixed vocal cord. One subglottic growth that was staged as a T, was found, at examination of serial sections, to extend paratracheally; therefore, this growth became a P4. Glottic and supraglottic cancer that extended onto the arytenoid cartilage was studied for invasion into the cartilage itself, and for its relation to such clinical features as vocal cord fixation and extent of visible mucosal spread. The total numbers of such cases in each category were too small to allow for conclusions. However, our impression from these few cases, is that vocal cord fixation in the presence of clinical involvement of the arytenoid is usually associated with deep invasion of the thyroarytenoid muscle; it is this latter feature that is responsible for this fixation. Assessment of invasion of the laryngeal framework is difficult. Xeroradiography is often helpful, but laryngeal tomography is more reli¬ able. Lateral laryngography is useful in evaluating the anterior commissure in the presence of cancer at the base of the epiglottis. It may also reveal preepiglottic space invasion by show¬ ing posterior displacement of the epiglottis. Roentgenographic demon¬ stration of the ventricle is sometimes misleading: a bulky lesion of the superior surface of the vocal cord may mimic a ventricular or small, trans¬ glottic cancer. The ventricle is best examined at direct laryngoscopy by pushing aside the lower edge of the ventricular band or by pulling it upward on a hook. Submucosal tumor in the base of the tongue is difficult to evaluate without palpation under general anesthesia.
COMMENT
The 40% rate of error in staging resulted mainly from a failure to predict extension of cancer outside the confines of the larynx. Therefore, there should have been more T4 and fewer T3 lesions in the original clinical staging. However, these errors might not affect the choice of a surgical procedure in supraglottic cancer, where the preepiglottic space is removed, as completely with horizon¬ tal supraglottic partial laryngectomy, as with total laryngectomy. However, with invasion into the base of the tongue, understaging could be disas¬ trous.
transglottic lesions, failure to anticipate extension of cancer through the thyroid or cricoid carti¬ lages, or through the cricothyroid membrane, could lead to inadequate surgical resection (Fig 3 and 4). Simi¬ larly, understaging may result in poor control rates when radiotherapy is the sole modality. With
CONCLUSIONS Present staging systems fail to identify invasion and destruction of the laryngeal framework in a high percentage of cases. This feature accounts for most of the 40% error in predicting the growth patterns of laryngeal cancer. Since tumor volume
presents
a
greater therapeutic prob¬
lem to the radiotherapist than to the surgeon, unsuspected deep invasion of tumor may account for relatively poorer control rates by radiotherapy than by surgery for any particular
stage. References 1. Tucker GF Jr: A histological method for the study of the spread of carcinoma within the larynx. Ann Otol 70:910-921, 1961. 2. TMN Classification of Malignant Tumors, ed 2. Geneva, Union Internationale Contre de Cancer, 1974. 3. Norris CM, Kuo BF, Tucker GF Jr, et al: A correlation of clinical staging, pathological findings and five-year end results in surgically treated cancer of the larynx. Ann Otol Rhinol Laryngol 79:1033-1048, 1970. 4. Kirchner JA, Som ML: Clinical and histological observations on supraglottic cancer. Ann Otol Rhinol Laryngol 80:638-645, 1971. 5. Kirchner JA, Cornog JL Jr, Holmes RE: Transglottic cancer: Its growth and spread within the larynx. Arch Otolaryngol 99:247-251, 1974.
Downloaded From: http://archotol.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/29/2015
